Seahorse Maritime Ltd v Nautilus International

The term “establishment” in section 188 of the 1992 Act derives from the EU Collective Redundancies Directive (98/59/EC) (or, strictly, its predecessor (75/129/EEC)). It is also used in other EU employment protection legislation and the UK legislation implementing it. Its meaning has proved problematic and has been considered in a number of domestic and EU authorities. I should start with the EU case-law since this must be authoritative as far as it goes.

An important part of the background is that the relevant provision of the Directive, article 1 (1) (a), gives member states a choice of two methods for determining the minimum number of proposed redundancies that will generate consultation obligations: both methods employ the term “establishment”. The methods are:

The first decision of the CJEU to consider the meaning of the term “establishment” was Rockfon A/S v Specialarbejderforbundet i Danmark (C-449/93), [1996] ICR 673. The relevant Danish legislation employed method (i). Four distinct production units were operated by subsidiaries of the same group. The employees at each unit were employed by the relevant subsidiary, but decisions about hiring and firing were taken by the group personnel department. Redundancies were carried out in one of the units only: the numbers affected did not fall within the first or third bullets but would fall within the second if each unit was a separate establishment. The employer argued that that was not the case, and that the entire organisation constituted a single establishment. The relevant passage from the Court’s judgment is at paras. 23-33, which read (pp. 688-9):

“23.

As regards the second part of the question, it must be noted first of all that the term ‘establishment’ is not defined in Directive (75/129/E.E.C.).

24.

Rockfon maintains that it is not an ‘establishment’ for the purposes of the Directive since it has no management which can independently effect large-scale dismissals and it does not therefore fulfil the condition, laid down by [the relevant legislation], for constituting an ‘establishment’. In its view, in counting the number of workers for the purposes of article 1(1)(a) of the Directive, all the workers of the four companies must be taken into account, not only the number of its own workers.

25.

The court observes in this regard that the term ‘establishment’, as used in Directive (75/129/E.E.C.), is a term of Community law and cannot be defined by reference to the laws of the member states.

26.

The various language versions of the Directive use somewhat different terms to convey the concept in question: the Danish version has ‘virksomhed,’ the Dutch version ‘plaatselijke eenheid,’ the English version ‘establishment,’ the Finnish version ‘yritys,’ the French version ‘etablissement,’ the German version ‘Betrieb,’ the

Greek version ‘επιχειρηση’ the Italian version ‘stabilimento,’ the Portuguese version ‘estabelecimento,’ the Spanish version ‘centro de trabajo’ and the Swedish version ‘arbetsplats.’

27.

A comparison of the terms used shows that they have different connotations signifying, according to the version in question, establishment, undertaking, work centre, local unit or place of work.

28.

As was held in Reg v Bouchereau [1978] Q.B. 732, 758, para. 14, the different language versions of a Community text must be given a uniform interpretation and in the case of divergence between the versions the provision in question must therefore be interpreted by reference to the purpose and general scheme of the rules of which it forms part.

29.

Directive (75/129/E.E.C.) was adopted on the basis of articles 100 and 117 of the E.E.C. Treaty, the latter provision concerning the need for the member states to promote improved working conditions and an improved standard of living for workers, so as to make possible their harmonisation while the improvement is being maintained. It is apparent from the first recital in its preamble that the Directive is indeed intended to afford greater protection to workers in the event of collective redundancies.

30.

Two observations may be made in that respect. First, an interpretation of the term ‘establishment’ like that proposed by Rockfon would allow companies belonging to the same group to try to make it more difficult for Directive (75/129/E.E.C.) to apply to them by conferring on a separate decision-making body the power to take decisions concerning redundancies. By that means, they would be able to escape the obligation to follow certain procedures for the protection of workers and large groups of workers could be denied the right to be informed and consulted which they have as a matter of course under the Directive. Such an interpretation therefore appears to be incompatible with the aim of the Directive.

31.

Secondly, the court has held that an employment relationship is essentially characterised by the link existing between the employee and the part of the undertaking or business to which he is assigned to carry out his duties: see Botzen v. Rotterdamsche Droogdok Maatschappij B.V. (Case 186/83) [1985] E.C.R. 519, 528, para. 15.

32.

The term ‘establishment’ appearing in article 1(1)(a) of Directive (75/129/E.E.C.) must therefore be interpreted as designating, depending on the circumstances, the unit to which the workers made redundant are assigned to carry out their duties. It is not essential, in order for there to be an ‘establishment’, for the unit in question to be endowed with a management which can independently effect collective redundancies.

33.

That interpretation is supported by the fact that the Commission's initial proposal for a Directive used the term ‘undertaking’ and that that term was defined in the last sub-paragraph of article 1(1) of the proposal as ‘local employment unit’. It appears, however, that the Council decided to replace the term ‘undertaking’ by the term ‘establishment’, which meant that the definition originally contained in the proposal and considered to be superfluous was deleted.”

(The case of Botzen referred to in para. 31 concerns the transfer of undertakings, where the relevant Directive also refers to “establishment”.)

Although I have set out the whole of the passage because of its importance to the argument, the essential point is that the Court defined “establishment” as “the unit to which the workers made redundant are assigned to carry out their duties”, even if the authority to make redundancies lay at some higher level: see para. 32.

That approach was confirmed but amplified in Athinaïki Chartopoïïa AE v Panagiotidis (C-270/05), [2007] IRLR 284. The employer had three production units manufacturing different kinds of paper product, but the main business decisions were taken by its head office. The relevant Greek legislation likewise employed method (i), and, as in Rockfon, the application of the second bullet meant that it was in the employer’s interests to argue that the entire organisation was a single establishment. The Court rejected that argument. Having summarised some of the points made in Rockfon, it continued, at paras. 27-32 (pp. 286-7):

“27.

Thus, for the purposes of the application of Directive 98/59, an 'establishment', in the context of an undertaking, may consist of a distinct entity, having a certain degree of permanence and stability, which is assigned to perform one or more given tasks and which has a workforce, technical means and a certain organisational structure allowing for the accomplishment of those tasks.

28.

Given that the objective pursued by Directive 98/59 concerns, in particular, the socio-economic effects which collective redundancies may have in a given local context and social environment, the entity in question need not have any legal autonomy, nor need it have economic, financial, administrative or technological autonomy, in order to be regarded as an 'establishment'.

29.

It is, moreover, in this spirit that the Court has held that it is not essential, in order for there to be an 'establishment', for the unit in question to be endowed with a management which can independently effect collective redundancies (Rockfon, paragraph 34, and point 2 of the operative part). Nor must there be a geographical separation from the other units and facilities of the undertaking.

30.

In the light of those considerations, the Court finds, first of all, that the fact that Greek Law 1387/1983 uses the terms 'establishment' and 'operating unit' interchangeably is not in itself contrary to Directive 98/59, provided that the interpretation given by the Court of the concept of 'establishment' is followed and the use of two terms cannot lead to the exclusion of categories of workers from the protection intended by that Directive.

31.

Next, regarding the nature of the production unit at issue in the main proceedings, the information in the case-file indicates that that unit is one of three separate production units held by the company; it employs 420 workers; it has distinct equipment and a specialised workforce; its operation is not affected by that of the other units; and it has a chief production officer who ensures that the work is carried out properly, is responsible for supervision of the entire operation of the unit's installations and ensures that technical questions are solved.

32.

Those factors clearly give such a unit the air of an 'establishment' for the purposes of the application of Directive 98/59, in accordance with the considerations set out by the Court in paragraphs 27 to 29 of the present judgment, and bring the unit in question within the scope of this Community concept. The fact that decisions concerning the operating expenditure of each of those units, the purchase of materials and the costing of products are taken at the company's headquarters, where a joint accounts office is set up, is irrelevant in this regard.”

Unlike Denmark and Greece, the UK has adopted method (ii). It will be apparent that under that method the choice made in Rockfon in favour of – in shorthand – the work unit rather than the entire organisation will necessarily result in consultation obligations arising in fewer cases, since even where the employer is making over twenty redundancies in its organisation as a whole the numbers affected at any one unit may be below the threshold. In two more recent cases in the CJEU – Lyttle v Bluebird UK Bidco 2 Ltd (C-182/13), [2015] 3 CMLR 33, (to which I will refer as “Bluebird”) and USDAW v Ethel Austin Ltd (C80/14), [2015] ICR 675, (“USDAW”) (Footnote: ³) – the employers operated chains of retail stores, with the employees in question each working exclusively in a particular store and thus “assigned” to it. (In Bluebird the stores were the Bonmarché chain of stores. USDAW comprised two separate references, both in proceedings brought by USDAW, arising out of the closures of, respectively, the Ethel Austin and the Woolworths chains.) The employers argued, on the basis of the Rockfon definition, that they would not be obliged to consult about proposed redundancies in any case where the store employed under twenty employees. The employees’ representatives argued that in cases where method (ii) applied a different definition should be adopted.

The Court did not accept that argument. At paras. 47-51 of its judgment in USDAW (pp. 703-4) it recited the effect of its decisions in Rockfon and Athinaïki Chartopoïïa. Despite an element of repetition it is worth setting out most of that passage here:

“47.

In para 31 of the judgment in Rockfon, the court observed, referring to para 15 of the judgment in Botzen v Rotterdamsche Droogdok Maatschappij BV (Case 186/83) [1985] ECR 519, that an employment relationship is essentially characterised by the link existing between the worker and the part of the undertaking or business to which he is assigned to carry out his duties. The court therefore decided in Rockfon, para 32, that the term ‘establishment’ in article 1(1)(a) of Directive 98/59 must be interpreted as designating, depending on the circumstances, the unit to which the workers made redundant are assigned to carry out their duties. It is not essential in order for there to be an ‘establishment’ that the unit in question is endowed with a management that can independently effect collective redundancies.

48.

….

49.

In Athinaïki Chartopoiïa the court further clarified the term ‘establishment’, inter alia, by holding, in para 27, that, for the purposes of the application of Directive 98/59, an ‘establishment’, in the context of an undertaking, may consist of a distinct entity, having a certain degree of permanence and stability, which is assigned to perform one or more given tasks and which has a workforce, technical means and a certain organisational structure allowing for the accomplishment of those tasks.

50.

By the use of the words ‘distinct entity’ and ‘in the context of an undertaking’, the court clarified that the terms ‘undertaking’ and ‘establishment’ are different and that an establishment normally constitutes a part of an undertaking. That does not, however, preclude the establishment being the same as the undertaking where the undertaking does not have several distinct units.

51.

In Athinaïki Chartopoiïa, para 28, the court held that, since Directive 98/59 concerns the socio-economic effects that collective redundancies may have in a given local context and social environment, the entity in question need not have any legal autonomy, nor need it have economic, financial, administrative or technological autonomy, in order to be regarded as an ‘establishment’.”

The Court then went on to hold, at paras. 53-67 (pp. 704-6), that that approach applied equally where the member state in question had opted for method (ii). I need not set out the reasoning in full, but I should note that it confronted expressly an argument that such a result would limit worker protection, contrary to the purposive approach adopted at paras. 29-30 of the judgment in Rockfon. As to that, it observed at para. 62 (p. 705D) that

“the objective of [the] Directive is not only to afford greater protection to workers in the event of collective redundancies, but also to ensure comparable protection for workers’ rights in the different member states and to harmonise the costs which such protective rules entail for EU undertakings”.

That reflects rather fuller observations to essentially the same effect in the Opinion of Advocate-General Wahl. At paras. 46- 47 (p. 667 D-F), he said:

“46.

… [O]ne of the lessons learnt from Rockfon and Athinaïki Chartopoiïa is that the court does not pay any heed to the way in which the employer-entity is structured internally, focusing instead on the local employment unit: see, in particular, Rockfon at para 30, and Athinaïki Chartopoiïa at para 28. To change stance now because an employer has several local employment units with fewer than 20 employees would, unlike before, pave the way for a malleable construction of that concept dependent on the employer’s internal structure, and that in turn would be at odds with recital (11) in the Preamble to Directive 98/59.

47.

Indeed, the approach argued for by the applicants in Cases C182/13 and C-80/14 is to extend the protective procedure to all workers dismissed in the course of the same restructuring exercise, irrespective of the size of the establishment at which they worked. Conferring the maximum level of protection by downplaying the method of implementation would obviously be to the advantage of those workers who, under the current understanding of the concept at issue, are not entitled to any protective award. However, such an approach would not be consonant with the minimum harmonisation aim envisaged by Directive 98/59, which, as the Commission rightly stated at the hearing, does not contemplate as a starting point full protection for all—even where the number of dismissals exceeds the thresholds— as the temporal requirement must also be met.”

The Court, as is its way, did not draw a definitive conclusion that the stores in question were distinct establishments, respecting the formal position that that was a matter for the national court. But the references were plainly made on the basis that if the Rockfon definition applied that would be the case, and in Bluebird the Court did go so far as to say, at paras. 51-52 of its judgment:

“51.

In the present case, on the basis of the information available to the Court …, it appears that each of the stores at issue in the main proceedings is a distinct entity that is ordinarily permanent, entrusted with performing specified tasks, namely primarily the sale of goods, and which has, to that end, several workers, technical means and an organisational structure in that the store is an individual cost centre managed by a manager.

52.

Accordingly, such a store is capable of satisfying the criteria set out in the case-law cited … above relating to the term ‘establishment’ in Article 1(1)(a) of Directive 98/59 … .”

We were referred by counsel to two domestic authorities, though both pre-date Bluebird/USDAW. I take them in turn.

In City of Edinburgh Council v Wilkinson [2011] CSIH 70, [2012] IRLR 202, the claimants were female employees of the council who had brought an equal pay claim. The issue was whether they were employed “at the same establishment”, within the meaning of section 1 (6) of the Equal Pay Act 1970, as their male comparators. In deciding what constituted an “establishment” the Inner House had regard to the CJEU authorities on the meaning of that term in the context of the transfer of undertakings and collective redundancies – specifically to Botzen, Rockfon and Athinaïki Chartopoïïa. At paras. 21-22 of his opinion (p. 207) Lord Essie, with whom Lady Paton and Lord Hardie agreed, observed that, while organisational questions might be relevant, “the term ‘establishment’ is largely directed to the place of work”. Also at

In Renfrewshire Council v Educational Institute of Scotland [2012] UKEAT 0018/12, [2013] ICR 172, the Council had made redundant over twenty teachers across the schools for which it was responsible but the numbers at any one school were below the threshold. It contended that each school was a separate establishment, to which individual teachers were assigned. The ET did not accept that argument but held that the relevant establishment was “the Education and Leisure Service”, which was (it seems) an arm of the Council. The EAT (Langstaff P sitting alone) overturned that decision, applying Rockfon and Athinaïki Chartopoïïa, and remitted the case to the ET. Much of the reasoning focuses on particular errors made by the ET, but I should note four points.

First, USDAW and Bluebird not having at that point been decided, Langstaff P was troubled by whether the definition of “establishment” adopted in Rockfon and Athinaïki Chartopoïïa should be extended to method (ii) cases. He was influenced in particular by the adoption in Rockfon of an approach avowedly designed to advance the Directive’s purpose of worker protection, which the Rockfon reading would not do in method (ii) cases. In the end he decided – in line with what the CJEU itself eventually held – that the language used could not mean different things in different parts of the same article of the Directive. But in the context of that discussion he said, at para. 26 (p. 182 C-E):

“… [W]here a choice has to be made on the facts between holding that either a greater or lesser unit is the relevant ‘establishment’, in circumstances where it would be permissible logically to conclude that either or both were the unit to which the relevant number of workers were assigned to carry out their duties, it is a relevant consideration in choosing between them that one choice will afford early consultation rights to those workers, whilst the other will not. [Counsel for the employer] in argument conceded as much, though he had at first been disinclined to do so … .”

Secondly, the ET had attached weight, as counting against “assignment”, to the fact that although all or most of the teachers worked at particular schools their contracts had mobility clauses under which they could be required to work elsewhere. Langstaff P, referring to Wilkinson, held that that had been a misdirection, since “‘assignment’ does not refer to the contractual, but rather to the factual, position” – see para. 42 (p. 185 G-H).

Thirdly, the ET had also attached weight to the fact that control of HR functions was exercised by the Council, not individual schools. That was of course directly contrary to the reasoning in Rockfon, and Langstaff P again held that it was a misdirection. He said, at para. 45 (p. 187B):

“Contrary to [Lady Smith’s] view, actual assignment, rather than power to control it, is the relevant criterion …; and it is to be expected, rather than being of signal note, that an undertaking will centrally control much of recruitment and possibly dismissal.”

Fourthly, Langstaff P was initially inclined to the view that it was unnecessary to remit the case because the only possible answer was that the relevant establishment in each case was the school. Though in the end he did remit, it was only because of three very particular features of the way in which the ET had set about its factfinding: see paras. 65-66 (p. 191 B-G). Among other things, it had not focused on how many of the employees whose redundancy was proposed were in fact assigned to the schools in question: this could be important because it was arguable that there was a class of peripatetic teachers who were not assigned to any particular school, and their inclusion might affect whether the relevant threshold was met. In describing the task of the ET on remittal he said, at para. 66 (p. 191 E-F):

“The central question ought to be capable of relatively simple resolution … and ought to focus on the individual employees rather than council powers: the question is employee, not employer, focused. It is as to which unit the worker is assigned to perform his duties. That involves two central questions—is the postulated unit capable of being an establishment; and, if so, is the employee assigned to it?”

It will be convenient to say here, though it anticipates one of the disputed points in the appeal, that the ET and the EAT may have treated the passage which I have set out in the previous paragraph as of wider import than Langstaff P intended. He was addressing a particular error which the ET had made and which he wished to make clear should be avoided on the remittal. The two questions which he identifies, and his references to the focus being on the individual employees, must be read in that context. At the risk of spelling things out unnecessarily, where there is an issue about whether the threshold in section 188 (1) has been satisfied, the following questions may arise:

The ET’s direction on the law as regards the establishment issue is at paras. 37-38 of the Reasons. It essentially consists of references to the case-law, and I need not set it out here.

The ET’s dispositive reasoning is at paras. 44-48 of the Reasons, which read as follows:

“44.

The Respondent contends that each of ships operated by its client (Sealion) is the establishment to which the Respondent’s employees were assigned to carry out their duties. This is not what the wording of the contracts of employment suggests, nor does it appear to be reflected in the way in which the redundancy exercise was carried out by FMA on the Respondent’s behalf.

45.

Each ship may be an individual business unit within Sealion’s business (as opposed to the Respondent’s business) – that is not irrelevant – but it does not necessarily follow that each ship is the establishment at which the Respondent proposed to dismiss a number of employees. The Respondent does not control where the ships work, what work they do or whether they are working or not. Generically individual ships are capable of being establishments, but in this case it cannot really be said that each one of the ships operated by Sealion are distinct parts of the Respondent’s undertaking.

46.

The authorities suggest a focus on the employee rather than the employer. Most employees went to and from the same ship on their 4 or 6 week roster for periods of time but, in accordance with their contracts, which did not state any particular ship as their place of work, some employees were transferred to other ships (such as those who had been on the Conqueror and the Envoy) – and others (the riding crew), were not considered to be attached to any particular ship. The employees’ duties commenced when they left their homes, and included the period of travel to the ships. Aside from payroll, which was dealt with in Guernsey, the employees turned to FMA in

Farnham in relation to any administrative queries. In some of the

correspondence from FMA on behalf of the Respondent, relating to the run up to the redundancies, the employees appeared to be treated as a group rather than only in relation to each individual ship (in particular the correspondence of 27 July 2015).

47.

The classic collective redundancy situation will involve an employer making redundancies at the employer’s establishment (e.g. a Woolworths store). That is certainly at least suggested by the wording of s188(4)(c) TULR(C)A 1992 ‘employed by the employer at the establishment’. The Tribunal agrees with the Respondent that it does not necessarily follow that the legislation excludes the possibility of the ‘establishment’ being outside the control of the employer and the Claimant puts its cases too highly in that respect. Section 188 does not have the words ‘of the employer’ after the word ‘establishment’ – however the fact that the ships are not establishments of the Respondent is one factor weighing against the Respondent’s arguments. I also bear in mind the comment of Mr Justice Langstaff in Renfrewshire at paragraph 26 as set out above.

48.

Both sides have good points to make but on balance, the Tribunal considered that the Respondent’s arguments concentrated on demonstrating that the individual ships operated by Sealion were distinct from each other rather than examining how and where the employees were assigned to. The Tribunal preferred the Claimant’s argument that the reality reflected the contractual situation that the employees were assigned to any of the vessels operated by Sealion. (Footnote: ⁴) That part of the Respondent’s workforce on each ship – forming some but not all of the crew of each ship – did not on the facts before the Tribunal, fulfil the USDAW/Rockfon/Athinaiki definition of ‘establishment’.”

I need to elucidate the references in para. 44 to “the way in which the redundancy exercise was carried out by FMA on the Respondent’s behalf” and in para. 46 to “the correspondence of 27 July 2015”. On that date FMA wrote to “all ships”, copied to

“all Seahorse crew” at their personal addresses, attaching a three-page “general information update to all Seahorse employees”. The document describes the general background to the redundancy situation and how redundancy selection is to be made. The particular passage which the ET evidently had in mind reads:

“There is apparently some confusion as to exactly who is at risk at present. The redundancy consultation affects all Seahorse employees – Guernsey and Auckland. … Crew may well be transferred where their abilities and experience can best be used. Just because ‘their’ ship is laid up does not put them more at risk than others. Alternative ships, positions, etc, will be considered wherever possible … .”

Seahorse’s grounds of appeal as advanced by Mr Pilgerstorfer to the EAT were essentially twofold (though this is not quite how they were pleaded):

Slade J addressed those grounds at paras. 37-53 of her judgment. Given that we are ultimately concerned with the reasoning of the ET I can, without disrespect to her, summarise her reasoning quite shortly. She identifies, in reliance on the passage from Langstaff P’s judgment in the Renfrewshire case quoted at para. 29 above, two distinct questions – (1) “whether the postulated unit is capable of being an establishment”, which she describes as “the establishment question”; and (2) “if so, whether the employee is assigned to it” (“the assignment question”). She says at para. 50 that the ET had to some extent conflated the two questions, but she holds that that did not vitiate its conclusion on the establishment question. As I read it, though this is not quite explicit, she took the view that the only possible conclusion on the facts found, and the evidence, was that the individual ships were not establishments, because the facts did not “show that the Respondent organised their business into individual ship based units” (para. 49). She says at para. 46 that a relevant factor in reaching that conclusion was that the ships themselves were not operated by Seahorse. She appears to acknowledge that, as Mr Pilgerstorfer had emphasised, the evidence tended to show that crews were in fact typically assigned to single ships, contrary to the finding at para. 48 of the Reasons, which indeed she described at para. 53 of her judgment as an “error”. But on her analysis that point, being relevant only to the assignment question, went nowhere in the light of her conclusion on the establishment question.

Mr Cavanagh subjected paras. 44-48 of the ET’s Reasons to a searching analysis and submitted that they contained a number of errors of law. In summary:

It was Mr Cavanagh’s case that if the ET had addressed the correct question the only possible answer on its findings of primary fact was that each ship constituted a separate establishment of Seahorse, just as the individual stores did in Bluebird and USDAW. Typically most Seahorse employees were, and remained, assigned to particular ships for long periods. It was irrelevant that there was the contractual right to assign them to different ships, or that that might occur from time to time: the ship was still, in the language of the CJEU, “the unit to which … workers … are assigned to carry out their duties”. He relied on what Lord Essie said about mobility clauses in Wilkinson and on Langstaff P’s adoption of it in Renfrewshire: see paras. 24 and 27 above. A ship was clearly a distinct place of work, which Lord Essie identified as the central concept: see para. 24. It is of its nature, in the language of the Court in

Mr Stone in response contended that the essential flaw in Mr Cavanagh’s criticisms was that they conflated the question whether each ship was an establishment of Sealion’s with the question whether it was an establishment of Seahorse’s. In the case of all the authorities on which Mr Cavanagh relied the putative establishment was operated by the employer of the workforce. If that had been so in this case it might indeed be said that a ship was a distinct entity satisfying the Athinaïki Chartopoïïa criteria. But it was not. We were here concerned with an employer who was a labour-only supplier, which was a fundamentally different situation. Although such an employer might in principle organise its workforce in a way that mirrored its client’s organisation, it did not follow that it had to do so, and there was no evidence that that had occurred in the present case. Although individual employees might typically be supplied to work on specific ships for long periods, there was no evidence that Seahorse itself had discrete “employment units” for each ship. On the contrary, the apparent model was of a single undifferentiated workforce: he referred to para. 51 of the judgment in USDAW, where the Court made clear that the employer’s entire organisation might constitute an establishment “where the undertaking does not have several distinct units”. There was no Seahorse management or organisation on each ship: the entire management was the responsibility of Sealion.

Mr Stone pointed out that the Judge was plainly alive to this point. He referred to it expressly in para. 47 of the Reasons, where he described it as “one factor weighing against the Respondent’s arguments” (while rejecting the submission that it was a knock-out blow). It also lay behind the statement in the final sentence of para. 45 that while individual ships were capable of being establishments Sealion’s ships could not be said to be “distinct parts of [Seahorse’s] undertaking” (my emphases); and behind the final sentence of para. 48. Likewise it was what Slade J meant by saying that Seahorse had not shown “that [it] organised their business into individual ship based units”. This was of course the supposed error of law challenged by Mr Cavanagh as his point (5), but it was not an error and the ET and the EAT were entirely right to regard it as important.

In that context, Mr Stone submitted, most of the points to which Mr Cavanagh said that the ET was wrong to have regard were in fact legitimate. The “non-shipspecific” terms of the contract reflected the reality; and it was indeed relevant that the redundancy arrangements were regarded as applying across the fleet, since it confirmed that Seahorse’s own business was not organised ship-by-ship.

Mr Stone submitted that there was accordingly no error of law in the ET’s approach. He endorsed Slade J’s conclusion that on the evidence the only possible finding was that the ships did not constitute separate establishments to which Seahorse’s employees were assigned. But he submitted that in any event the question was one of fact and the ET had reached a conclusion which was undoubtedly open to it.

Well though Mr Stone developed that case, I cannot accept it. In none of the cases discussed above do the workers in question appear to have been employed by a different legal person than the operator of the establishment (Footnote: ⁵), but in my view it is sufficiently clear from the approach taken in the authorities, and in particular the CJEU authorities, that the identity of the employer is not in itself a relevant factor. The starting-point is that section 188 (1) does not say “at one establishment of his” or anything to the same effect. More substantially, the reasoning in all three of the judgments from which I have quoted focuses on functional and organisational characteristics – essentially whether the putative establishment constitutes a unit – and also, which is related, whether it is single “place”. As regards the latter point, I note not only Lord Essie’s observation in Wilkinson quoted at para. 24 above but also para. 28 of the judgment in Athinaïki Chartopoïïa (para. 18). In that context I cannot see that it is material whether the owner or operator of the unit is the same legal person as the employer of some or all of the workforce. It is not uncommon in groups of companies for the staff to be employed by a specific subsidiary which makes their services available to the different operating subsidiaries in the group; likewise, employees of different subsidiaries may constitute an integrated workforce at a particular unit. There is no reason why idiosyncratic complications of that kind should govern whether a particular unit constitutes an establishment. It is enough in my view that the establishment should have a workforce assigned to it (together with the other elements in the definition), irrespective of the identity of the employer or employers.

I should make it clear that that reasoning is specific to the present issue and the circumstances of the present case. In some cases the fact that a group of workers working at a particular unit are employed by a different person than the owner or operator of that unit may reflect the fact that they constitute a workforce not assigned to it. They might, for example, be a self-contained team moving from location to location on a short-term basis. If that were so, the unit in question might not fall to be regarded as an establishment at all, but even if it was the fact that the employees in question were not assigned to it would require the threshold to be applied by reference to their employer’s own organisation. But this is not such a case. (Footnote: ⁶)

On that basis, I believe that the correct conclusion on the Tribunal’s findings of primary fact is that each ship was indeed an establishment. It is clearly a selfcontained operating unit of the kind described in the case-law. The only possible question is whether it can be said to have a workforce assigned to it. As to that, the only evidence was that relating to the crew members supplied by Seahorse. The Tribunal concludes at para. 48 of its Reasons that Seahorse crew were not assigned to particular ships. That finding is, however, vitiated by the error discussed above, and also by the further errors identified, I believe correctly, by Mr Cavanagh which I have listed as points (1)-(4) at para. 36 above. I do not believe that it is (even ignoring the “territorial jurisdiction” issue) necessary to remit the issue to the Tribunal. It is in my view adequately clear from its findings of primary fact that, at least in the typical case, Seahorse crew were assigned to particular ships. The finding at para. 13 of the Reasons, repeated at para. 46, is that most employees returned, rota after rota, to the same ship for “periods of time”, and the Tribunal at para. 46 expressly contrasts that with the case of the “riding squad/crew”, who are not “considered to be attached to any particular ship”. A view that crew are generally attached to a particular ship is reflected also in the reference in the letter of 27 July 2015 to “‘their’ ship”: see para. 33 above. I also think that the appearance on the standard-form contract of the box

“Ship Name (if known)” suggests an expectation that in principle an employee will be attached to a particular ship, even if that has not been identified at the moment of contracting.

That conclusion does not address the position of any non-assigned employees, such as the riding squad/crew, and there are insufficient findings for me to express any conclusion about to what establishment they might be regarded as assigned and thus as to whether an obligation to consult might arise in relation to them.

My present understanding is that it is very likely that the effect of my conclusion as regards crew assigned to particular ships is that no obligation to consult arose. But that has not been definitively established, and for that reason it is necessary to consider the territorial jurisdiction issue also. I would in any event wish to do so since the point is of some importance.

“TERRITORIAL JURISDICTION”

As appears from the facts summarised at para. 8 above, the ships relevant to this appeal on which the Seahorse employees represented by Nautilus work – and which I have held to constitute the establishments to which they were (typically) assigned – were stationed outside Great Britain. The question accordingly arises of whether the rights conferred by Chapter II of Part IV of the 1992 Act extend to Nautilus and/or to the employees themselves. Strictly speaking, this is not so much – as it is expressed in the preliminary issue – a question of “jurisdiction” as of the territorial scope of the substantive rights (Footnote: ⁷); but nothing turns on that.

The 1992 Act contains no provision defining the territorial reach of the rights conferred by Chapter II of Part IV. There is the same lacuna in the Employment Rights Act 1996 and the Equality Act 2010, which confer a range of individual rights on workers against their employers. There were originally such provisions in both statutes – for that applying to the 1992 Act see para. 56 below – but they were repealed by section 32 (1) of the Employment Act 1999. There are no decided cases

It was common ground in the ET and the EAT, as also before us, that some form of the “sufficient connection” test must also be applied in the case of the rights under the 1992 Act with which we are concerned here. But the parties differ as to what the connection should be with. Mr Cavanagh submits that the question should be answered by reference to the establishment; and that in that case the only answer can be that there is no sufficient connection between the relevant ships and Great Britain. Mr Stone submits that the question should, rather, focus on the individual employees assigned to the establishment; and that a sufficient connection exists in the case of those Seahorse employees who live in Great Britain (whose numbers remain to be established – see para. 12 above) (Footnote: ⁸). Mr Stone accepts that there is no such connection in the case of those who live outside Great Britain.

The ET accepted Nautilus’s submission that the focus of the relevant enquiry should be on the employee. That being its approach, it found that the UK-based employees could properly be regarded as falling into the so-called “peripatetic employee exception” recognised by Lord Hoffmann in Lawson v Serco.

In the EAT Slade J upheld the ET’s decision that the focus of the sufficient connection test should be on the connection with the employee rather than with the establishment: I will come back to her reasons later. She too accordingly performed a conventional Lawson/Ravat analysis. At para. 93 of her judgment she held that on the primary facts found the ET’s decision that the UK-based Seahorse employees were peripatetic was perverse. However she went on in paras. 94-99 to accept an alternative submission by Mr Stone that they were “international commuters” of the kind identified in Ravat, and so had a sufficient connection with Great Britain to overcome the fact that they worked abroad.

The issues accordingly are:

ESTABLISHMENT OR INDIVIDUAL EMPLOYEES ?

The ET’s reasoning on this aspect is not easy to analyse, and since the issue is one of law I need not undertake the task. I should note, however, that the Judge appears to have relied, at least to some extent, on the decision of Supperstone J in Netjets Management Ltd v Central Arbitration Committee [2012] EWHC 2685 (Admin), [2012] IRLR 986, in which he upheld a decision of the CAC that a group of pilots employed by a UK company which operated a fleet of business jets operating within Europe constituted a “bargaining unit” for the purpose of the compulsory recognition procedure provided by Schedule A1 to the 1992 Act, notwithstanding that most of the pilots were not UK-based. Slade J held that he was wrong to do so, because the statutory context was different; and Mr Stone accepted that Netjets was of no assistance on this point.

As for the reasoning of the EAT, Slade J acknowledged at para. 79 of her judgment that the issue was not straightforward. Her reasons for taking an approach that focused on the connection with the individual, which Mr Stone adopted, fall under essentially two heads.

First, she attached importance to the fact that, although the right conferred by section 188 (1) itself is collective in character – that is, it is a right conferred on the employees’ representatives – the ultimate beneficiaries of the right are the employees themselves. If the duty to consult under section 188 is breached, the tribunal, as noted above, makes a protective award, by which a payment is required to be made to each of the employees affected: see section 189 (3). And if the employer fails to pay any sum due under a protective award the employees in question may bring proceedings in their own name: see section 192. Slade J says, at para. 87:

“Whilst it may be said that the obligations and entitlements provided by TULR(C)A ss 188 to 192 are hybrid rather than all collective or all individual, in my judgment [the] territorial jurisdictional reach of all such provisions is to be determined by the rights and means of enforcement which are given to the individual employees.”

Secondly, she referred to the legislative history. Section 285 of the 1992 Act, as originally enacted, reads (so far as material) as follows:

“Employment outside Great Britain

(1)

The following provisions of this Act do not apply to employment where under his contract of employment an employee works, or in the case of a prospective employee would ordinarily work, outside Great Britain—

In Part III …;

In Part IV, Chapter II (procedure for handling redundancies).

(2)

For the purposes of subsection (1) employment on board a ship registered in the United Kingdom shall be treated as employment where under his contract a person ordinarily works in Great Britain unless —

(a)

the ship is registered at a port outside Great Britain, or

(b)

the employment is wholly outside Great Britain, or

(c)

the employee or, as the case may be, the person seeking employment or seeking to avail himself of a service of an employment agency, is not ordinarily resident in Great Britain.”

Section 32 (1) of the 1999 Act amended section 285 (1), as regards Part IV, so that it applied only to two particular sections under Chapter II, namely sections 193 and 194 (which impose an obligation on an employer to notify the Secretary of State of proposed redundancies and provided for non-compliance to constitute an offence). It is that amendment which gives rise to the lacuna referred to above: section 285 (1) no longer has any application (and for that reason nor does sub-section (2)). At para. 86 of her judgment Slade J says:

“The legislative history of the exclusionary provisions of s 285 as they affect ss 188 to 192 demonstrated a focus on whether under his contract of employment an employee worked outside Great Britain rather than where the 'establishment' to which he is assigned is located. That express exclusion did not depend upon assignment to an establishment located outside Great Britain.”

Although I agree with Slade J that the issue is not easy, I have respectfully come to the opposite conclusion from her. My reasons, which largely follow Mr Cavanagh’s submissions, are as follows.

The starting-point is that the primary obligation created by section 188 is purely at the collective level. That is, the obligation is to consult with the employees’ representatives, normally a trade union, and not with individual employees (Footnote: ⁹) – albeit that if it is breached the remedy is by way of an obligation to make payments to individuals, enforceable by those individuals. That is not in any way decisive, but it means that the case is not straightforwardly comparable to those considered in the case-law referred to above, which concern rights conferred directly on individuals.

Given the collective character of the right, I think that, in asking whether it has a sufficient connection with Great Britain, and British employment law, the most natural focus is on the common feature which defines the group for whose benefit the right is created, namely that they are employed at the same establishment. Such a

First, different employees might have different degrees of connection with Great Britain. It is inherently unsatisfactory that, in order to decide whether consultation obligations are owed – and, if they are owed but are not complied with, which employees should be the beneficiaries of a protective award –individual assessments of the Lawson/Ravat kind should have to be carried out, initially by the employer but in the event of dispute by the tribunal, in every case. The assessment of whether there is a sufficient connection requires a careful balancing of the factors pulling in either direction and can be very fact-sensitive (see, for a recent example, the two cases considered in Jeffery (Footnote: ¹¹)). To take one particular factor, it is now established that the presence of an English choice of law clause in the contract of employment and/or any equivalent assurances given to the employee are relevant considerations (see Jeffery paras. 50-66).

Mr Stone did not dispute that that could be so in some cases, but he contended that in the present case all the Seahorse crew had identical contracts; and in any event it was his submission that in this case a sufficient connection would be established if the employee in question lives in Great Britain, which is on the face of it not difficult to establish. I am not persuaded that even on his case the exercise would be straightforward: in the case of a group of employees who, ex hypothesi, work abroad, there are likely to be a number where the question of whether they live in the UK is debatable. But the real point is that we are concerned with the question of the right general approach to the reach of the rights under section 188 (1), and the fact that in some cases an individual-focused enquiry would be straightforward is no answer to the fact that in other cases it would not be.

Secondly, if Nautilus’s approach is adopted the sufficient connection test may be satisfied in the case of some of the employer’s employees at the establishment in question and not others. Indeed that is its case here, since, as already noted, Mr Stone accepts that Seahorse employees who do not live in Great Britain do not fall within

Mr Stone says that there are equal practical difficulties with Mr Cavanagh’s approach.

As for Slade J’s point based on section 285 (1) in its original form, I do not think that the legislative history supports her conclusion. Although the original exclusion for

“employment where under his contract of employment an employee works … outside Great Britain” is indeed “individual-focused”, it would in practice have meant that no consultation obligations would be imposed as regards redundancies at establishments located abroad, and the kinds of difficulty that I consider in the previous paragraphs would not have arisen. In those circumstances it affords no guidance as to whether an individual-focused approach is appropriate in the fundamentally different situation created by the 1999 Act. Mr Cavanagh reminded us Lord Hoffmann’s observations in Lawson about the very limited extent to which the formulation of the predecessor provisions could shed any light on the position post-1999: see paras. 9-14 (pp. 255-7).

Mr Stone made a different point about section 285, namely that the “working outside Great Britain” exclusion remains applicable to sections 193 and 194. But essentially the same point arises. The fact that this “individual-focused” formulation is used in the context of the obligation to notify the Secretary of State about proposed redundancies casts no light on how to approach the obligation under section 188 (1).

IS THERE A SUFFICIENT CONNECTION IN THE PRESENT CASE ?

It follows from my answer to (1) that this question falls to be answered by reference to the establishment – that is, as I have held above, the individual ships – and not to the individual Seahorse employees assigned to them. Because they answered question (1) differently, neither the ET nor the EAT addressed this question. It was Mr Cavanagh’s submission that the answer was clear. It cannot have been the intention of Parliament, or the makers of the Directive, that employers should be obliged by British, or EU, law to consult about the making of redundancies at establishments on

Mr Stone’s case did not make any submissions on this variant of the issue: his case on “territorial jurisdiction” was directed to question (1). In those circumstances I need only say that there is nothing in the circumstances of the present case that would overcome the territorial pull of the location of the establishment. The only connections with Great Britain are that some (though not all) of Seahorse’s functions are performed through its agent, FMA, which is based in Farnham. But in the case of the 1996 and 2010 Acts the fact that the employer is based in Great Britain is not by itself enough to overcome the territorial pull of the place of work, and the present case seems to me analogous, if not indeed a fortiori.

We heard argument on whether, if the right approach was “individual-focused”, the EAT’s conclusion on the sufficient connection question – and specifically that the employees were “international commuters” of the kind recognised in Ravat – was correct. In that connection we were referred to the decisions of the EAT in Diggins v Condor Marine Crewing Services Ltd [2009] EWCA Civ 1133, [2010] ICR 213, and Windstar Management Services Ltd v Harris [2016] UKEAT 0001/16, [2016] ICR 847, which are both unfair dismissal cases. But it is unnecessary for us to consider that question. The issue on this appeal is about the rights under Chapter II of Part IV of the 1992 Act and it has no bearing on claims under the 1996 and 2010 Acts.

For those reasons I would allow Seahorse’s appeal and dismiss Nautilus’s claim as regards the proposed redundancies of any Seahorse crew assigned to any ships in the TOISA fleet stationed outside Great Britain.