ON APPEAL FROM THE EMPLOYMENT APPEAL TRIBUNAL
Mr Justice Singh
UKEAT/0189/15/BA
Royal Courts of Justice
Strand, London, WC2A 2LL
Before :
THE MASTER OF THE ROLLS
LADY JUSTICE GLOSTER
and
SIR COLIN RIMER
Between:
BRITISH GAS TRADING LIMITED | Appellant |
- and - | |
(1) MR Z. J. LOCK (2) SECRETARY OF STATE FOR BUSINESS, INNOVATION AND SKILLS | Respondents |
John Cavanagh QC (instructed by Eversheds LLP) for the Appellant, British Gas Trading Limited
Michael Ford QC and Simon Cheetham (instructed by Camilla Belch, Legal Officer, UNISON) for the First Respondent, Mr Z.J. Lock
Adam Tolley QC (instructed by the Government Legal Department) for the Second Respondent, the Secretary of State
Hearing date: 11 July 2016
Judgment
Sir Colin Rimer:
Introduction
This appeal is by British Gas Trading Limited (‘British Gas’). It is against an order of the Employment Appeal Tribunal (‘EAT’) dated 22 February 2016 made by Singh J, sitting alone, dismissing British Gas’s appeal against a judgment of the Leicester Employment Tribunal (Employment Judge Ahmed, sitting alone) (‘the ET’) dated 23 March 2015 and sent, with written reasons, to the parties on 25 March. The sole issue before the tribunals was one of statutory interpretation. The practical question that is raised is whether the holiday pay of an employee with statutorily defined ‘normal working hours’, whose remuneration does not vary with the amount of work done during such hours, should (i) be calculated solely by reference to his basic pay; or (ii) include an element referable to the amount of the results-based commission he normally earned.
On 10 April 2012, Mr Lock, a British Gas employee (respondent to the appeal), presented a claim in the ET complaining that British Gas had wrongly failed to calculate his holiday pay so as to include such a commission element. The complaint was formally one of alleged unauthorised deduction from his wages contrary to section 13 of the Employment Rights Act 1996. There is no dispute that, as a matter of European Union (‘EU’) law, article 7 of the applicable Directive required a commission element to be included. The dispute is as to whether the Working Time Regulations 1998, the UK legislation enacted to give domestic effect to the Directive, can be interpreted as incorporating that requirement.
The ET’s decision was that they can, although in so deciding it found it necessary to read words into them whose apparent effect is to amend the provision applying to Mr Lock’s case. Singh J, in the EAT, upheld that decision. Because of the importance of the point, he gave permission to British Gas for this further appeal. We were told that some 918 like claims against British Gas and thousands of like claims against other employers have been stayed pending its outcome. We had able arguments from Mr Cavanagh QC, for British Gas; Mr Ford QC and Mr Cheetham, for Mr Lock; and Mr Tolley QC, for the Secretary of State for Business, Innovation and Skills, who supported Mr Lock’s case.
The facts
These are agreed and I take them from the EAT’s judgment. Mr Lock became employed by British Gas on 1 February 2010 although he has since moved to other employment. His role was to persuade business customers to purchase British Gas’s energy products. At the material time, his annual salary was £14,670 but he was also entitled to the benefit of a commission scheme. He was entitled to 25 days’ holiday per year, plus public and bank holidays. Whilst on leave, he was paid at the rate of his basic salary of £1,222.50 per month. His claim related to the pay he received for the leave he took from 19 to 30 December 2011 and for the statutory holidays on 26/27 December 2011 and 2 January 2012: he complained that British Gas failed to include in it a commission element. He alleged an unlawful underpayment of £1,500.
The commission scheme of which Mr Lock was a beneficiary was designed to ‘provide an incentive to encourage and reward individual performance’ if, and only if, sales he had negotiated had ‘gone live’, that is to say, the customer had begun to take gas from British Gas. There were three methods of achieving a sale: (i) ‘cold calling’; (ii) ‘hot leads’, namely where a potential customer had already been contacted by British Gas, the conversion rate in such cases being much higher; and (iii) ‘upgrades’, which involved existing customers not under contract, but who are persuaded to enter into one, the conversion rate in such cases being even higher. The scheme was contained in a document separate from his employment contract but it is agreed that Mr Lock’s entitlement under it was a contractual right. His entitlement was based purely on success, namely a sale negotiated by him resulting in the customer beginning to take energy products from British Gas. As Singh J explained:
‘14. … the amount of work done by him in normal working hours did not vary in the sense that the payment was not based on the amount of work done. Rather, payment of commission was based on the outcome of that work, whether or not it was due to good performance. The amount of the commission was not based on the amount of work he carried out during any particular period. It was simply dependent on the outcome of his work: that is the number and type of new contracts which customers entered into.’
The commission that Mr Lock earned greatly exceeded his basic salary. Whilst on leave, he was paid any commission earned in previous periods that fell due for payment during his leave; but he says that his leave pay should have been calculated so as also to include a commission element: his argument is that as his remuneration normally included commission, so should his leave pay.
The legislation
The relevant EU legislation was originally in Council Directive 93/104/EC (‘the Working Time Directive’), enacted on 23 November 1993 and amended in 2000 by Council Directive 2000/34/EC (known as the Horizontal Amending Directive). The current EU Directive (‘the Directive’), a consolidation of those Directives, is Council Directive 2003/88/EC ‘concerning certain aspects of the organisation of working time.’ It was enacted on 4 November 2003 and came into force on 2 August 2004. Article 7 (‘Annual Leave’) provides:
‘1. Member States shall take the measures necessary to ensure that every worker is entitled to paid annual leave of at least four weeks in accordance with the conditions for entitlement to, and granting of, such leave laid down by national legislation and/or practice.
The minimum period of paid annual leave may not be replaced by an allowance in lieu, except where the employment relationship is terminated.’
The Working Time Directive was given national effect by the Working Time Regulations 1998 (SI 1998/1833) (‘the WTR’), which also gave effect to Council Directive 94/33/EC on the protection of young people at work. The WTR were made under section 2(2) of the European Communities Act 1972 and came into force on 1 October 1998. They have since been amended, the latest version dating from 1 October 2013. The relevant regulations are 13 and 16.
Regulation 13 (‘Entitlement to annual leave’) provides in paragraph (1) that, subject to paragraph (5) (which is not material), a worker is entitled to four weeks’ annual leave in each leave year. Regulation 16 (‘Payment in respect of periods of leave’), which is the key provision, provides:
‘(1) A worker is entitled to be paid in respect of any period of annual leave to which he is entitled under Regulation 13 … at the rate of a week’s pay in respect of each week of leave.
Sections 221 to 224 of the [Employment Rights Act 1996] shall apply for the purpose of determining the amount of a week’s pay for the purposes of this Regulation, subject to the modifications set out in paragraph (3).
The provisions referred to in paragraph (2) shall apply –
as if references to the employee were to the worker;
as if references to the employee’s contract of employment were references to the worker’s contract;
as if the calculation date were the first date of the period of leave in question; and
as if the reference to sections 227 and 228 did not apply.
A right to payment under paragraph (1) does not affect any right of a worker to remuneration under his contract (“contractual remuneration”) (and paragraph (1) does not confer a right under that contract).
Any contractual remuneration paid to a worker in respect of a period of leave goes towards discharging any liability of the employer to make payments under this Regulation in respect of that period; and, conversely, any payment of remuneration under this Regulation in respect of a period goes towards discharging any liability of the employer to pay contractual remuneration in respect of that period.’
The provisions of the Employment Rights Act 1996 (‘the ERA’) relating to the determination of the amount of ‘a week’s pay’, and incorporated into the WTR by regulation 16(2), are in Chapter II of Part XIV, headed ‘A week’s pay’. They date from the Contracts of Employment Act 1963 and are used for many statutory employment law purposes, for example the calculation of redundancy payments. Section 220 (‘Introductory’) provides that the amount of a week’s pay of an employee shall be calculated for the purposes of the ERA in accordance with Chapter II. Sections 221 to 223 are in a division of Chapter II headed ‘Employments with normal working hours’, which is Mr Lock’s case. Section 224 is in a division headed ‘Employments with no normal hours’, which is not his case. Sections 221 and 222 provide:
‘221. General
This section and sections 222 and 223 apply where there are normal working hours for the employee when employed under the contract of employment in force on the calculation date.
Subject to section 222, if the employee’s remuneration for employment in normal working hours (whether by the hour or week or other period) does not vary with the amount of work done in the period, the amount of a week’s pay is the amount which is payable by the employer under the contract of employment in force on the calculation date if the employee works throughout his normal working hours in a week.
Subject to section 222, if the employee’s remuneration for employment in normal working hours (whether by the hour or week or other period) does vary with the amount of work done in the period, the amount of a week’s pay is the amount of remuneration for the number of normal working hours in a week calculated at the average hourly rate of remuneration payable by the employer to the employee in respect of the twelve weeks ending –
where the calculation date is the last day of a week, with that week, and
otherwise, with the last complete week before the calculation date.
In this section references to remuneration varying with the amount of work done includes remuneration which may include commission or similar payment which varies in amount.
This section is subject to sections 227 and 228.’
As regards section 221(5), regulation 16(3)(d) provided for sections 221 to 224 to apply as if the references in it to sections 227 and 228 did not apply. Section 227 imposes a cap on the measure of ‘a week’s pay’ for various employment purposes (for example, in calculating awards for unfair dismissal or redundancy payments). Section 228 applies to ‘new employments and other special cases’ and need not be further described. I return to the incorporated provisions of the ERA:
‘222. Remuneration varying according to time of work
This section applies if the employee is required under the contract of employment in force on the calculation date to work during normal working hours on days of the week, or at times of the day, which differ from week to week or over a longer period so that the remuneration payable for, or apportionable to, any week varies according to the incidence of those days or times.
The amount of a week’s pay is the amount of remuneration for the average number of weekly normal working hours at the average hourly rate of remuneration.
For the purposes of subsection (2) –
the average number of weekly hours is calculated by dividing by 12 the total number of the employee’s normal working hours during the relevant period of 12 weeks, and
the average hourly rate of remuneration is the average hourly rate of remuneration payable by the employer to the employee in respect of the relevant period of 12 weeks.
For the purposes of subsection (3) “the relevant period of 12 weeks” means the period of 12 weeks ending –
where the calculation date is the last day of a week, with that week, and
otherwise, with the last complete week before the calculation date.
This section is subject to sections 227 and 228.’
As regards section 222(5), I make the like comment as in para 11 above. Section 223 (‘Supplementary’) is not material. Nor, for reasons given (see para 10 above), is section 224. Section 234 (‘Normal working hours’) is not expressly incorporated by regulation 16(2) but in Bamsey and others v. Albon Engineering and Manufacturing plc [2004] EWCA Civ 359; [2004] ICR 1083 this court decided that it did incorporate it. It provides:
‘(1) Where an employee is entitled to overtime pay when employed for more than the fixed number of hours in a week or other period, there are for the purposes of this Act normal working hours in his case.
Subject to subsection (3) the normal working hours in such a case are a fixed number of hours.
Where in such a case –
the contract of employment fixes the number, or minimum number of hours of employment in a week or other period (whether or not it also provides for the reduction of that number or minimum in certain circumstances), and
that number or minimum number of hours exceeds the number of hours without overtime,
the normal working hours are that number or maximum number of hours (and not the number of hours without overtime).’
The reference to the Court of Justice of the European Union
There is no dispute that if Mr Lock’s entitlement under the WTR to holiday pay is assessed through a domestic legislative lens, his case is covered by section 221(2) of the ERA, with the consequence that his pay is confined to his basic salary and excludes any commission element. The length of the journey from the presentation of Mr Lock’s claim in 2012 to the appeal in this court in July 2016 is in part explained by the fact that on 16 November 2012 the ET stayed the proceedings and made a reference to the CJEU for a preliminary ruling. The ruling sought was as to whether, in the case of a worker like Mr Lock whose remuneration normally consists of a basic salary and commission, with the amount of the latter being fixed by reference to contracts entered into as a result of sales he achieves, article 7 of the Directive limits his holiday pay to remuneration composed only of his basic salary.
By its judgment of 22 May 2014 (Lock v. British Gas Trading Ltd (Case C-539/12); [2014] ICR 813, the CJEU (First Chamber) held that that is not the effect of article 7. As to Mr Lock’s salary and commission arrangements, the court said:
‘10. For December 2011, his remuneration was composed of the basic pay of £1,222.50 and commission which he had earned over previous weeks amounting to £2,350.31. In 2011, Mr Lock earned on average monthly commission of £1,912.67.
Given that Mr Lock did not carry out any work during his period of annual leave, he was not able to make any new sales or follow up on potential sales during that period. Accordingly, he was not able to generate commission during that period. Since this had adverse effects on the salary Mr Lock received during the months following his annual leave, he decided to bring an action before the referring tribunal for outstanding holiday pay in respect of the period from 19 December 2011 to 3 January 2012.’
The court re-affirmed its case law (dating from 2006) that the entitlement of every worker to paid annual leave was a particularly important principle of EU law from which there can be no derogation ‘and whose implementation by the competent national authorities must be confined within the limits expressly laid down by’ the Working Time Directive, the Directive and the court’s case law. It continued:
‘16. Although the wording of article 7 does not give any express indication as regards the remuneration to which a worker is entitled during his annual leave, the court has already stated that the term “paid annual leave” in article 7(1) means that, for the duration of “annual leave” within the meaning of that Directive, remuneration must be maintained and that, in other words, workers must receive their normal remuneration for that period of rest: see Robinson-Steele v. RD Retail Services Ltd (Joined Cases C-131/04 and C-257/04) [2006] ICR 932; [2006] ECR I-2531, para 50, and Stringer v. Revenue and Customs Comrs (Joined Cases C-350/06 and C-520/06) [2009] ICR 932; [2009] ECR I-179, para 58.
Directive 2003/88 treats entitlement to annual leave and to a payment on that account as being two aspects of a single right. The purpose of providing payment for that leave is to put the worker, during such leave, in a position which is, as regards his salary, comparable to periods of work: see Robinson-Steele, para 58, and Stringer, para 60.’
The court rejected British Gas’s case that Mr Lock did receive remuneration during his leave period comparable to what he earned during his periods of work. British Gas’s point was that his leave pay comprised not just his basic salary but also the commission arising during his leave period resulting from sales he had achieved during the weeks preceding his leave. The court appears to have recognised that that meant that Mr Lock’s leave pay was comparable to his remuneration earned during periods of work. But the flaw in the argument was, in its view, that he still faced the disadvantage that, whilst on leave and so not working, he was unable to generate commission that would be paid to him subsequently, an inability adversely affecting the remuneration he later received. That adverse consequence might, in the court’s view, deter a worker such as Mr Lock from actually taking his annual leave; and it noted that that was all the more likely in a case like Mr Lock’s whose results-based commission represented on average over 60% of his remuneration. The court said that such a reduction of remuneration in respect of annual leave, one liable to deter the worker from taking such leave, was contrary to the objective of article 7; and it referred to British Airways plc v. Williams (Case C-155/10) [2012] ICR 847; [2011] ECR I-8409, para 21.
The court’s conclusion was, therefore, that article 7(1) precluded national legislation and practice under which a worker whose remuneration was comprised of basic salary and results-based commission was entitled to holiday pay comprised exclusively of his basic salary. As to how the holiday pay of a worker like Mr Lock should be calculated, the answer was that it should correspond to his ‘normal remuneration’; and, where such remuneration is composed of several components, ‘the determination of the normal remuneration to which the worker in question is entitled during his annual leave requires a specific analysis’, and the court referred again to para 21 of Williams. It thus made it clear that the commission received by a worker such as Mr Lock must be taken into account in calculating the pay in respect of his annual leave; but it left it to the national court to work out the commission element to which he was so entitled.
There is, therefore, no dispute that Mr Lock’s holiday pay was not calculated in accordance with what article 7 required and that, as a matter of EU law, it ought, in addition to his basic salary, to have included an element referable to the commission he ordinarily earned when working. Only then would his holiday pay be comparable to his ‘normal remuneration’.
I add that it is agreed that the CJEU’s ruling in Lock applies only to the four weeks of annual leave provided for by regulation 13 of the WTR. It does not apply also to the extra 1.6 weeks of annual leave provided for, as a matter of UK domestic law, by regulation 13A (which does not derive from the Directive), nor does it apply to any additional contractual leave period.
The decision of the ET
Following the decision of the CJEU, the case returned to the ET, which of course recognised Mr Lock’s entitlement as a matter of EU law. His problem, however, was that it was said against him by British Gas that, whilst the sense of article 7 was clear, the WTR giving domestic effect to the Directive not only do not provide for results-based commission to be taken into account when determining the amount of his holiday pay, their interpretation shows it is not to be taken into account. That was a crucial point since, although a Directive which has not been correctly or fully transposed into national law may be directly relied upon by an individual against the member state concerned, or against state bodies, the same does not apply in relation to an individual’s claim against a private sector entity such as British Gas. If Mr Lock was to succeed he had, therefore, to show that the WTR could be interpreted in conformity with article 7 as now explained by the EU.
None of that was in dispute before the ET. What was in dispute was whether the WTR can be so interpreted. The ET described the issue as being:
‘2. … whether … domestic law, in the form of the [WTR], can be read consistently with EU law and if not whether words can and should be added in interpreting those regulations so that the calculation of a week’s pay is in conformity with EU law.’
The ET explained that the scheme of sections 221 to 224 of the ERA distinguished between workers with normal working hours (to whom sections 221 and 222 apply) and those who do not (to whom section 224 applies). It was agreed that: (i) section 221(3), which applied where the employee’s remuneration for his normal hours of work ‘varied with the amount of work done in the period’ was not Mr Lock’s case; (ii) nor did section 222 apply, that section applying to shift workers, which was also not his case; (iii) nor did section 224 apply, since he did have normal working hours. It was and is agreed that the only provision applying to Mr Lock’s case is section 221(2).
After reviewing recent decisions of the CJEU, including the Lock case, the ET considered the domestic law. It referred to this court’s decision in Evans v. The Malley Organisation Ltd (t/a First Business Support) [2003] ICR 432, where the material facts were indistinguishable from Mr Lock’s case and the court held that, as the case fell within section 221(2), there was no question of commission payments being brought into account in the calculation of holiday pay. The court gave no consideration to whether its interpretation was consistent with the requirements of article 7. That is not surprising: the decision pre-dated the first CJEU case (in 2006) explaining that ‘paid annual leave’ in article 7(1) means that, for the duration of annual leave within the meaning of the Directive, a worker’s ‘normal remuneration’ must be maintained: see Robinson-Steele v. RD Retail Services Ltd (Case C-131/04) [2006] ICR 932, at para 50. The ET accepted that:
‘42. … having regard to the Court of Appeal’s judgment in Evans, domestic legislation cannot, in accordance with the normal rules of statutory interpretation, be read so as to require employers to take results-based commission payment into account when calculating pay for annual leave.’
It nevertheless rejected the submission that the interpretation applied to section 221(2) in Evans rendered impermissible an interpretation of the WTR in conformity with the requirements of article 7 as since understood.
The ET referred next to the EAT’s decision in Bear Scotland Ltd v. Fulton and another [2015] ICR 221 as to whether ‘non-guaranteed’ overtime had to be included in leave pay for the purposes of article 7 (that is, overtime the employer can require the employee to work but which it has no duty to offer). Having held it did, the EAT inserted additional words into the WTR having the apparent effect of amending them so as bring them into line with the requirements of article 7 of the Directive. That decision was cited to the ET as decided by a ratio said to apply also to Mr Lock’s case. The ET then engaged in a review of the authorities and of the parties’ submissions, observing en route that (a) most of British Gas’s arguments had been run and rejected in Bear Scotland and (b) that the remaining arguments were without substance.
The ET concluded (in essence) that the issue in Bear Scotland was no different in principle from that in Mr Lock’s case; that it accepted Bear Scotland’s answers to British Gas’s objections against a conforming interpretation of the WTR; that it was permissible and necessary to imply words into the WTR ‘for it to comply with EU law’; that such an exercise did not involve going against the grain of the legislation or its underlying thrust; that it would do no violence to the intention of Parliament, nor would it amount to ‘judicial vandalism’ (a phrase used by Lord Bingham of Cornhill in R (Anderson) v. Secretary of State for the Home Department [2003] 1 AC 837, para 30); rather, it would fulfil the duty of interpretation; the ET was satisfied that, in enacting the WTR, Parliament’s intention was to comply with the Directive.
The ET accepted, however, that the language of the WTR (and the provisions of the ERA it incorporates) cannot be read in a way conforming with article 7 without reading words into it enabling it to be so read. After considering four suggested alternatives, the ET’s judgment adopted this wording:
‘Regulation 16(3) of the Working Time Regulations 1998 is to be interpreted and applied as if it had the following paragraph added to it:
as if, in the case of the entitlement under Regulation 13, a worker with normal working hours whose remuneration includes commission or similar payment shall be deemed to have remuneration which varies with the amount of work done for the purpose of section 221.’
The effect of the new para (e) is to deem a ‘commission or similar payment’ case that, but for such paragraph, would fall within section 221(2) as instead falling within section 221(3) and so to bring into play the 12-week averaging exercise for which it provides. We were told that, following the parties’ subsequent agreement, the ET delivered a separate judgment recording expressly that the reference periods for the calculation of Mr Lock’s ‘normal remuneration’ should be the same periods as provided by section 221(3).
The decision of the EAT
I can take this shortly. In support of British Gas’s appeal against the ET’s decision, it was submitted (i) that Bear Scotland was distinguishable; alternatively (ii) that Bear Scotland wrongly ignored what was said to be this court’s binding decision in Bamsey [2004] EWCA Civ 359; [2004] ICR 1083 that the WTR were not capable of the claimed conforming interpretation; alternatively (iii) that Bear Scotland should anyway not be followed. Singh J held that Bear Scotland was not distinguishable. He considered whether, as it was a decision of the EAT, he could or should depart from it. In his view, the only bases on which he might or should do so were if it was ‘manifestly wrong’ or there were ‘exceptional circumstances’ justifying it. He held that neither case applied and dismissed the appeal.
The appeal to this court
The EAT’s decision in Bear Scotland is not binding on this court and I shall come to what it decided. Although the EAT granted the employers in that case permission to appeal, no appeal was pursued. The heart of Mr Cavanagh’s submission was that Bear Scotland was wrongly decided and that this case is not one in which it was open to the ET, as he said it did, to amend the WTR under the guise of interpretation. He submitted in the alternative, albeit with express moderation, that Bamsey anyway binds us to allow the appeal.
The rest of this judgment follows this course. I shall: (i) refer to the principal authorities on ‘conforming interpretation’; (ii) explain the decision in Evans; (iii) do likewise in relation to the decisions in Bamsey; and (iv) Bear Scotland; (v) summarise the submissions; and (vi) express my conclusions.
Authorities on conforming interpretation
The ECJ’s decision in Marleasing S.A. v. LA Comercial Internacional de Alimentacion S.A. (Case C–106/89) [1992] 1 CMLR 305 is a well known authority on ‘conforming interpretation’. The ECJ said, at para 8:
‘… It follows that, in applying national law, whether the provisions in question were adopted before or after the directive, the national court called upon to interpret it is required to do so, so far as possible, in the light of the wording and the purpose of the directive in order to achieve the result pursued by the latter and thereby comply with the third paragraph of Article 189 EEC.’
Ghaidan v. Godin-Mendoza [2004] UKHL 30; [2004] 2 AC 557 is a decision of the House of Lords on the application of section 3 of the Human Rights Act 1998, which requires that ‘[s]o far as it is possible to do so, primary and subordinate legislation must be read and given effect in a way which is compatible with Convention rights.’ It was agreed before us that the principles applicable to the section 3 interpretative obligation are the same as those applicable to the Marleasing interpretative obligation: in his speech in Ghaidan, Lord Steyn noted, at para 45, that the draftsman of section 3 had resorted to the analogy of that obligation.
The question in Ghaidan was whether the surviving homosexual partner of the deceased protected tenant of a flat succeeded to his tenancy as a statutory tenant by succession as his surviving ‘spouse’ within the meaning of paragraph 2 of Schedule 1 to the Rent Act 1977, which extended the word ‘spouse’ to persons living with the deceased tenant ‘as his or her wife or husband’. The House held that, unless the legislation could be read as extending to the partner’s case, his Convention rights were violated. There thus arose the question whether section 3 enabled a Convention-compliant interpretation to be applied to Schedule 1. Upholding the Court of Appeal, the House held (Lord Millett dissenting) that it could.
Lord Nicholls of Birkenhead noted that the operation of section 3 does not depend on there being an ambiguity in the legislation in question: even if there is no doubt as to its meaning according to the ordinary principles of interpretation, section 3 may nonetheless require it to be given a different meaning. He gave this guidance:
‘30. From this it follows that the interpretative obligation decreed by section 3 is of an unusual and far-reaching character. Section 3 may require a court to depart from the unambiguous meaning the legislation would otherwise bear. In the ordinary course the interpretation of legislation involves seeking the intention reasonably to be attributed to Parliament in using the language in question. Section 3 may require the court to depart from this legislative intention, that is, depart from the intention of the Parliament which enacted the legislation. The question of difficulty is how far, and in what circumstances, section 3 requires the court to depart from the intention of the enacting Parliament. The answer to this question depends upon the intention reasonably to be attributed to Parliament in enacting section 3.
On this the first point to be considered is how far, when enacting section 3, Parliament intended that the actual language of a statute, as distinct from the concept expressed in that language, should be determinative. Since section 3 relates to the “interpretation” of legislation, it is natural to focus attention initially on the language used in the legislative provision being considered. But once it is accepted that section 3 may require legislation to bear a meaning which departs from the unambiguous meaning the legislation would otherwise bear, it becomes impossible to suppose Parliament intended that the operation of section 3 should depend critically upon the particular form of words adopted by the parliamentary draftsman in the statutory provision under consideration. That would make the application of section 3 something of a semantic lottery. If the draftsman chose to express the concept being enacted in one form of words, section 3 would be available to achieve Convention-compliance. If he chose a different form of words, section 3 would be impotent.
From this the conclusion which seems inescapable is that the mere fact the language under consideration is inconsistent with a Convention-compliant meaning does not of itself make a Convention-compliant interpretation under section 3 impossible. Section 3 enables language to be interpreted restrictively or expansively. But section 3 goes further than this. It is also apt to require a court to read in words which change the meaning of the enacted legislation, so as to make it Convention-compliant. In other words, the intention of Parliament in enacting section 3 was that, to an extent bounded only by what is “possible”, a court can modify the meaning, and hence the effect, of primary and secondary legislation.
Parliament, however, cannot have intended that in the discharge of this extended interpretative function the courts should adopt a meaning inconsistent with a fundamental feature of legislation. That would be to cross the constitutional boundary section 3 seeks to demarcate and preserve. Parliament has retained the right to enact legislation in terms which are not Convention-compliant. The meaning imported by application of section 3 must be compatible with the underlying thrust of the legislation being construed. Words implied must, in the phrase of my noble and learned friend, Lord Rodger of Earlsferry “go with the grain of the legislation”. Nor can Parliament have intended that section 3 should require courts to make decisions for which they are not equipped. There may be several ways of making a provision Convention-compliant, and the choice may involve issues calling for legislative deliberation.’
Lord Rodger’s concurring speech contains valuable observations on the extent and limits of the section 3 interpretative power. He devotes material parts to illustrating the difference between the permissible interpretation of a statute by an exercise under that power and the impermissible amendment of it by a purported such exercise (see paras 110–113). From those paragraphs, I shall simply cite the following extract from para 111, which discussed the decision of the House of Lords in R (Anderson) v. Secretary of State for the Home Department [2003] 1 AC 837:
‘111. … In these circumstances, in the words of Lord Bingham of Cornhill [2003] 1 AC 837, 883, para 30:
“To read section 29 as precluding participation by the Home Secretary, if it were possible to do so, would not be judicial interpretation, but judicial vandalism: it would give the sections an effect quite different from that which Parliament intended and would go well beyond any interpretative process sanctioned by section 3 of the 1998 Act …”
The “judicial vandalism” would lie not in any linguistic changes, whether great or small, which the court might make in interpreting section 29 but in the fact that any reading of section 29 which negatived the explicit power of the Secretary of State to decide on the release date for murderers would be as drastic as changing black into white. It would remove the very core and essence, the “pith and substance” of the measure that Parliament had enacted – to use the familiar phrase of Lord Watson (in a different context) in Union Colliery Co of British Columbia Ltd v. Bryden [1899] AC 580, 587. Section 3(1) gives the courts no power to go that far. …’
Lord Rodger’s observations in the following paragraphs (referring to authorities on the Marleasing interpretative principle) are also important:
‘121. For present purposes it is sufficient to notice that cases such as Pickstone v. Freeman plc [1989] AC 66 and Litster v. Forth Dry Dock & Engineering Co Ltd [1990] 1 AC 546 suggest that, in terms of section 3(1) of the 1998 Act, it is possible for the courts to supply by implication words that are appropriate to ensure that legislation is read in a way which is compatible with Convention rights. When the court spells out the words that are to be implied, it may look as if it is “amending” the legislation, but that is not the case. If the court implies words that are consistent with the scheme of the legislation but necessary to make it compatible with Convention rights, it is simply performing the duty which Parliament has imposed on it and on others. It is reading the legislation in a way that draws out the full implications of its terms and of the Convention rights. And, by its very nature, an implication will go with the grain of the legislation. By contrast, using a Convention right to read in words that are inconsistent with the scheme of the legislation or with its essential principles as disclosed by its provisions does not involve any form of interpretation, by implication or otherwise. It falls on the wrong side of the boundary between interpretation and amendment of the statute.
When Housman addressed the meeting of the Classical Association in Cambridge in 1921, he reminded them that the key to the sound emendation of a corrupt text does not lie in altering the text by changing one letter rather than by supplying half a dozen words. The key is that the emendation must start from a careful consideration of the writer’s thought. Similarly, the key to what it is possible for the courts to imply into legislation without crossing the border from interpretation to amendment does not lie in the number of words that have to be read in. The key lies in a careful consideration of the essential principles and scope of the legislation being interpreted. If the insertion of one word contradicts those principles or goes beyond the scope of the legislation, it amounts to impermissible amendment. On the other hand, if the implication of a dozen words leaves the essential principles and scope of the legislation intact but allows it to be read in a way which is compatible with Convention rights, the implication is a legitimate exercise of the powers conferred by section 3(1). Of course, the greater the extent of the proposed implication, the greater the need to make sure that the court is not going beyond the scheme of the legislation and embarking upon amendment. Nevertheless, what matters is not the number of words but their effect. For this reason, in the Community law context, judges have rightly been concerned with the effect of any proposed implication, but have been relaxed about its exact form. See, for example, Lord Keith of Kinkel and Lord Oliver of Aylmerton in Pickstone v. Freemans plc [1989] AC 66, 112D, 126A–B.’
Lord Millett dissented on the outcome of the case, but we were referred to what he said at para 68:
‘In my view section 3 does not entitle the court to supply words which are inconsistent with a fundamental feature of the legislative scheme; nor to repeal, delete or contradict the language of the offending statute. …’
I would not regard the first part of that, down to ‘legislative scheme’, as out of line with what others said in Ghaidan. There may be a question as to whether, by comparison, his remaining observations favoured too narrow an approach.
Vodafone 2 v. Revenue and Customs Commissioners [2009] EWCA Civ 446; [2010] Ch 77 is a decision of this court in which Sir Andrew Morritt C set out the following summary of the principles of conforming interpretation:
‘37. We were referred in the parties’ respective written arguments and orally to a number of reported cases on the principles to be observed in looking for a conforming interpretation in either the European Community or Human Rights contexts. In chronological order they are Pickstone v. Freemans plc [1989] AC 66; Marleasing SA v. La Comercial Internacional de Alimentacion SA [1990] ECR I-4135; Litster v. Forth Dry Dock & Engineering Co Ltd [1990] 1 AC 546; Imperial Chemical Industries plc v. Colmer (No 2) [1999] 1 WLR 2035; Ghaidan v. Godin-Mendoza [2004 2 AC 557; R (IDT Card Services Ireland Ltd) v. Customs and Excise Comrs [2006] STC 1252; Revenue and Customs Comrs v. EB Central Services Ltd [2008] STC 2209 and the Fleming/Condé Nast cases [2008] 1 WLR 195. The principles which those cases established or illustrated were helpfully summarised by counsel for HMRC in terms from which counsel for V2 did not dissent. Such principles are that:
“In summary, the obligation on the English courts to construe domestic legislation consistently with Community law obligations is both broad and far-reaching. In particular: (a) it is not constrained by conventional rules of construction (per Lord Oliver of Aylmerton in the Pickstone case, at p. 126B); it does not require ambiguity in the legislative language (per Lord Oliver in the Pickstone case, at p. 126B and Lord Nicholls of Birkenhead in Ghaidan’s case, at para 32); (c) it is not an exercise in semantics or linguistics (per Lord Nicholls in Ghaidan’s case, at paras 31 and 35; per Lord Steyn, at paras 48–49; per Lord Rodger of Earlsferry, at paras 110–115); (d) it permits departure from the strict and literal application of the words which the legislature has elected to use (per Lord Oliver in the Litster case, at p 577A; per Lord Nicholls in Ghaidan’s case, at para 31); (e) it permits the implication of words necessary to comply with Community law obligations (per Lord Templeman in the Pickstone case, at pp 120H–121A; per Lord Oliver in the Litster case, at p 577A); and (f) the precise form of the words to be implied does not matter (per Lord Keith of Kinkel in the Pickstone case, at p 112D; per Lord Rodger in Ghaidan’s case, at para 122; per Arden LJ in the IDT Card Services case, at para 114).”
Counsel for HMRC went on to point out, again without dissent from counsel for V2, that:
“The only constraints on the broad and far-reaching nature of the interpretative obligation are that: (a) the meaning should ‘go with the grain of the legislation’ and be compatible with the underlying thrust of the legislation being construed’: see per Lord Nicholls in Ghaidan v. Godin-Medoza [2004] 2 AC 557, para 53; Dyson LJ in Revenue and Customs v. EB Central Services Ltd [2008] STC 2209, para 81. An interpretation should not be adopted which is inconsistent with a fundamental or cardinal feature of the legislation since this would cross the boundary between interpretation and amendment (see per Lord Nicholls, at para 33, Lord Rodger, at paras 110–113 in Ghaidan’s case; per Arden LJ in R (IDT Card Services Ireland Ltd) v. Customs and Excise Comrs [2006] STC 1252, paras 82 and 113); and (b) the exercise of the interpretative obligation cannot require the courts to make decisions for which they are not equipped or give rise to important practical repercussions which the court is not equipped to evaluate: see the Ghaidan case, per Lord Nicholls, at para 33; per Lord Rodger, at para 115; per Arden LJ in the IDT Card Services case, at para 113.’
In Swift (trading as A Swift Move) v. Robertson [2014] UKSC 50; [2014] 1 WLR 3438, the Supreme Court, in a unanimous judgment, said this:
‘20. A national court must interpret domestic legislation, so far as possible, in the light of the wording and purpose of the Directive which it seeks to implement. This is now well settled. Thus in Schulte v. Deutsche Bausparkasse Badenia AG (Case C-350/03) [2003] All ER (EC) 420, para 71 the CJEU said:
“when hearing a case between individuals, the national court is required, when applying the provisions of domestic law adopted for the purpose of transposing obligations laid down by a Directive, to consider the whole body of rules of national law and to interpret them, so far as possible, in the light of the wording and purpose of the Directive in order to achieve an outcome consistent with the objective pursued by the Directive (see Pfeiffer v. Deutsches Rotes Kreuz, Kreisverband Waldshut eV (Joined Cases C-397/01 to C-403/01) [2005] ICR 1307, para 120).”
The breadth and importance of the principle was authoritatively set out in Vodafone 2 v. Revenue and Customs Comrs [2010] Ch 77 …
and the court cited the material parts of the passages I have cited.
Finally, in United States of America v. Nolan [2015] UKSC 63; [2015] ICR 1347, another decision of the Supreme Court, Lord Mance said, at para 14:
‘Taking the first point of construction, it is a cardinal principle of European and domestic law that domestic courts should construe domestic legislation intended to give effect to a European Directive so far as possible (or so far as they can do so without going against the “grain” of the domestic legislation) consistently with the Directive ….’
Evans v. The Malley Organisation Ltd
Evans ([2003] ICR 432) explains why an interpretation of the WTR exclusively through a domestic lens mandates a calculation of Mr Lock’s holiday pay confined to his basic pay and excluding any element referable to the commission he normally earned as a result of his successful efforts to induce customers to sign up to British Gas’s energy products.
The facts in Evans are not relevantly distinguishable from those of Mr Lock’s case. Mr Evans’s contract provided for his rate of holiday pay to be his ‘normal basic rate’, a reference to his basic salary of £10,000 a year. He was also entitled to commission under his contract, which arose and was payable in a like way as Mr Lock’s. His claim, also reliant upon the WTR, was that his holiday pay should have been his average salary (by reference to his basic salary and commission), not just his basic salary, which is all his employer paid him.
The ET dismissed his claim. It held that Mr Evans was an employee who was remunerated for employment in normal working hours, whose remuneration did not vary with the amount of work done in the period and whose case thus fell within section 221(2) of the ERA. It also referred to the fact that his contract provided for his holiday pay to be at the basic rate, although I do not understand that consideration to have been central to the ET’s decision.
On Mr Evans’s appeal, subject to the effect of section 221(4) of the ERA, the EAT agreed with the ET’s interpretation of sections 221(1) to (3). In delivering the EAT’s judgment, His Honour Judge Wilkie QC said, at para 12:
‘The amount of work done in the period of normal working hours did not vary in the sense that payments were not based on the amount of work done. Rather payment of commission was based on the outcome of that work, whether fortuitous or due to good performance. Therefore, as these three [subsections] stand, one would have thought that the natural meaning was that commission would not be included as part of the payment, so that the averaging out provisions of subsection (3) would not apply.’
The EAT then, however, departed from the ET by holding that section 221(4) operated to override the ordinary meaning of section 221(3) and so bring Mr Evans’s case within section 221(3) rather than section 221(2). The result was that the EAT allowed his appeal.
On the employer’s further appeal, this court unanimously disagreed with the EAT’s interpretation of section 221(4) and restored the ET’s decision. Pill LJ said:
‘23. I am unable to accept Mr Cohen’s submission. The distinction between subsection (2) and subsection (3) of section 221 turns on whether or not the employee’s remuneration does or does not vary with the amount of work done in the normal working hours. I am unable to conclude that it does. Work is done and the amount of work does not depend on the number of contracts obtained. Time spent unsuccessfully to persuade a client to sign a contract is as much work as a successful encounter with a client. I am not able to read the expression “amount of work done” as meaning that amount of work and that part of the work which achieves a contract. The amount of work resulting in a contract may vary, but the result achieved by the work is a different concept from the act of working.
In my judgment subsection (4) does not bear upon the issue whether a contract falls within subsection (2) or (3). That must first be determined in accordance with the test plainly stated in the section and already identified. Subsection (4) is not relevant to that decision. What subsection (4) and its predecessors achieve is to make clear that once the categorisation is made, the relevant remuneration may include commission or similar payments such as a bonus. It is not otiose because it is easy to envisage situations in which remuneration does vary with the amount of work done, once a specified level of productivity has been achieved. The reference to commission in subsection (4) does not require or permit all contracts in which commission is a part of the remuneration to be placed within subsection (3). …
What the use of the averaging method does tend to confirm, however, is the fit between subsection (3) and pieceworking in the traditional sense. Where there are marked variations in the amount of work done as between one week and another fairness can be achieved by calculating the amount of holiday pay by reference to an average. That objective is a fair but limited one. Its inclusion in the statute does not require that contracts such as the present should be forced into the subsection (3) category.’
Judge LJ said:
‘35. Mr Evans was of course expected to work conscientiously, and if he did it was hoped, both by him and his employers, that he would be successful in obtaining contracts. For these efforts he was paid his basic salary, which was due to him whether he succeeded in obtaining any contracts or none. If by working conscientiously he also achieved what it was hoped that he would achieve, he would then, but not otherwise, have earned commission in addition to his salary. Therefore the payment of commission did not depend on the length of his working week, and his remuneration for his employment was linked, not with the amount of work which he did, but with its success. Naturally it was hoped, indeed anticipated, that harder work and more skilful salesmanship would increase the number of contracts obtained by Mr Evans and so increase his resulting commission. But taken on their own, admirable though they are, hard work and skill which produced no contracts entitled him to no more than his basic salary.
For the purposes of section 221 of the Employment Rights Act 1996 Mr Evans’s remuneration did not vary with the amount of work he did during his working week. Any commission due to him was payable by virtue of earlier success, usually many months previously. It was unconnected with the amount of work he did during the 12-week period before his employment came to an end, which forms the basis of any calculation under section 221(3), and on which the decision of the Employment Appeal Tribunal was founded.
Unlike the appeal tribunal, I do not believe that section 221(4) overrides section 221(3) or, as seems to be implicit in its decision and was supported in argument by Mr Cohen, that cases in which commission forms part of a remuneration package must automatically be treated as falling with section 221(3). Rather, section 221(4) amplifies section 221 and, where remuneration does in fact vary with the amount of work done, enables commission and bonuses and similar payments to be included in the calculation of an employee’s week’s pay.’
Hale LJ said:
‘43. There are several good reasons to conclude that although this remuneration varied it did not vary “with the amount of work done”: (i) “work done” would ordinarily mean tasks undertaken, such as researching potential clients, making telephone calls, writing letters, meeting potential clients: it would not mean “success achieved”. Mr Cohen quite rightly says that work done leads to success achieved: but that does not mean that the words have the same meaning. (ii) The ordinary meaning of the “amount” of work done would refer to its quantity and not to its quality or its results. (iii) The variation in remuneration in this case was not “with” the amount of work done in the period but with success achieved as a result of work done in a completely different period, usually nine months earlier. (iv) The concept of averaging over 12 weeks is difficult to fit with the concept of success fees relating to a completely different period.’
After citing para 12 of the EAT’s judgment (see para 44 above), Hale LJ continued:
‘45. There is nothing in section 221(4) to change that. This is clearly defining remuneration for the purpose of what is included as remuneration but that still has to be within the overall criterion of varying with the amount of work done.’
The decision in Evans thus focussed solely on the interpretation of the relevant provisions of the WTR and their referential incorporation of the ERA as a matter of domestic law: it gave no consideration to whether such interpretation conformed with the requirements of article 7. I have explained why there was a good reason for that (see para 23 above). Whilst, however, the court did not recognise it, the court was in fact answering essentially the same question: namely, whether under the WTR and the incorporated provisions of the ERA, Mr Evans’s holiday pay should have been calculated by reference to his average (or normal) remuneration and not just his basic salary. That required the court to consider whether it was possible to construe the legislation as entitling him to what might be called a ‘normal remuneration’ calculation; and the court held it was not.
Evans shows to my satisfaction that it is not possible to interpret the WTR by conventional domestic canons of construction as entitling Mr Lock to holiday pay that includes any commission element. The ET also recognised that but it was concerned with what it rightly recognised as the different question of whether, in compliance with its duty to adopt so far as possible an interpretation of the WTR conforming with the CJEU’s post-Evans explanation of the meaning of article 7, it was able so to interpret the WTR. The ET held that it was, although only by reading into the WTR the words it did. The authorities show that a conforming interpretation may require the court to read words into the relevant legislation and that, subject to the qualifications they identify, the scope for doing so is very wide. The question in this appeal is whether it was permissible for the ET to interpret the WTR in the way it did.
Bamsey and others v. Albon Engineering and Manufacturing plc
Bamsey ([2004] ICR 1083), also a decision of this court, concerned a question similar to that later destined to arise in Bear Scotland, which declined to follow it. Mr Cavanagh submitted that Bamsey binds us to allow this appeal. Like Evans, Bamsey was decided before Robinson-Steele, the first CJEU decision to indicate that the holiday pay to which workers are entitled under article 7 means their ‘normal remuneration’.
Mr Sturge (the test case appellant in one of several appeals) was employed under a contract requiring him to work 39 hours a week and up to nine hours of overtime if required. His overtime was thus ‘non-guaranteed’, meaning he had to work it if required to do so, but that he was not entitled to be offered overtime. His holiday pay was fixed at the pay rate for his 39-hour week and included nothing reflecting the normal amount of overtime he had earned during the prior 12-week period. His claim that this breached regulation 16 of the WTR and article 7 of the Directive was dismissed by the ET. His appeal to the EAT failed as also did his further appeal to this court (Auld, May and Jacob LJJ).
A central question was the meaning of ‘normal working hours’ for the purposes of sections 221 to 224 of the ERA. They are defined in section 234, which, the court held, make clear that the hours of non-guaranteed overtime worked by Mr Sturge were not part of his ‘normal working hours’ for the purposes of the calculation of his holiday pay. There was, however, a dispute as to whether regulation 16 incorporated not just sections 221 to 224 of the ERA (to which it refers expressly) but also section 234 (to which it does not). It was agreed that, if regulation 16 did incorporate section 234, the appeals must fail.
The appellants’ argument was that articles 2 and 7 of the Directive required member states to ensure that a worker’s holiday pay was calculated by reference to their normal pay (a submission of notable prescience, the ECJ not having yet so decided); that regulation 16 should be purposively construed to achieve that result; that there was nothing in sections 221 to 224 of the ERA, read without section 234, to preclude that construction; and that, even if section 234 did govern sections 221 to 224 for the purposes of the ERA, it should, for various reasons, be ignored when construing regulation 16. The advocate tto the court’s contrary argument was that the Directive was silent on there being a minimum requirement as to the level of a worker’s holiday pay; there was therefore no basis for interpreting the ERA as if section 234 did not apply to sections 220 to 224; and even if the WTR were contrary to the Directive, they could anyway not be construed in a conforming way.
Auld LJ delivered the lead judgment, with which May and Jacob LJJ agreed. He held that, in a case where overtime was involved, section 234 was an essential aid to the determination of whether or not the hours of overtime formed part of the worker’s ‘normal working hours’; and that regulation 16 did incorporate section 234 even though it did not refer to it expressly.
Auld LJ also held that there was nothing in the Directive to require the purposive interpretation for which the appellants contended. That was because article 7 left it to ‘national legislation and/or practice’ to determine the conditions of entitlement to, and granting of, a worker’s paid annual leave, which necessarily included the ‘definition of the basis upon which payment is calculated for such period of leave’ and the level of such payment. He concluded that part of his discussion with these paragraphs:
‘39. … common sense also points to the conclusion that the Directive had to leave it to member states to decide how to calculate the amount of remuneration payable in respect of the absolute entitlement to four weeks’ paid annual leave. The pay systems of different employers across the European Union differ; a workable common definition would, therefore, be difficult to achieve. In the United Kingdom alone, sections 221-224 illustrate the range of issues that would need resolution, for example where pay varies according to the amount of work done, or time worked or where there are no normal working hours, questions as to what benefits are to be included in pay for this purpose and whether pay should be calculated at basic or enhanced overtime rates. It follows, in my view, that, unless the conditions of entitlement laid down by regulation 16, as I have construed it, are such that they can be said to negate or frustrate the very purpose of the Directive, the court must look at the regulations unassisted in this respect by the Directive.
In my view, there is nothing in regulation 16 on which the Marleasing [1990] ECR I-4135 principle of construction can bite, especially where, as I have concluded, the content and framework of the 1998 Regulations, when read with the 1996 Act, show that their draftsman clearly intended to apply the Act’s well established domestic definition of “a week’s pay” save in the immaterial respects for which he specifically provided in regulation 16(3). In particular, there is no basis for reading article 7 of the Directive as requiring a broad equivalence of pay for work done, namely overtime, which the employer was not bound to provide under the contract of employment, with payment on annual leave for overtime work not done at all. And, in any event, sections 221 to 224, with or without section 234, will not necessarily achieve that. As I have mentioned, section 223 is capable of producing in individual cases as “week’s pay” that may be more or less than an employee actually earned over the 12-week period.
Further, although the Directive was intended to have the effect of encouraging workers, for the sake of their health, to take their full leave entitlement, which they might not if their holiday pay is significantly less than their normal working pay, it could equally be said that it was not intended to encourage them to enter into contractual arrangements in which they submitted themselves to obligatory long and unhealthy working hours for 11 months of the year by the additional carrot of a level of holiday pay to match such hours when they were not actually working them. There was no evidence before the appeal tribunal or other basis upon which it could have found that employees would or might as a generality be discouraged from exercising their entitlement to paid annual leave where their weekly pay for the purpose is calculated by reference to section 234 of the Act.’
Mr Cavanagh relied upon para 40 in his submission that Bamsey binds us to hold that the WTR are not susceptible of an interpretation conforming with article 7 as explained by the CJEU. I shall return to this when summarising his submissions.
Bear Scotland Ltd v. Fulton and another
Had Bamsey come before the courts just a few years later, it would have arisen against a background in which the European landscape had changed significantly: namely, that contrary to the court’s unanimous view, article 7 in fact worked a secret magic that the Directive’s language does not betray – that member states do not have a free hand as to the basis of the calculation of holiday pay but must ensure that it at least matches a worker’s normal remuneration. That was decided by the various CJEU authorities that the CJEU cited in its judgment in Lock. Thus the premise upon which this court in Bamsey assessed the requirements of the Directive in relation to the determination of the level of holiday pay under the WTR proved to be false.
The like question then, however, arose in Bear Scotland ([2015] ICR 221). This decision lay at the heart of British Gas’s undoing before the ET and the EAT. It was a decision of the EAT, whose judgment was delivered by Langstaff J, the President, a judge with very considerable employment law experience. There were in fact two appeals before the EAT, each raising similar questions. The central issue was whether the remuneration in respect of the employees’ non-guaranteed overtime should have been (but was not) taken into account in calculating their holiday pay.
The EAT accepted that article 7 required non-guaranteed overtime to be paid during the employees’ annual leave. The next question was whether the ET had been right to find that the WTR could be interpreted in a way conforming with article 7 and so achieve that result. A material part of the EAT’s discussion of that addressed, and rejected, the employers’ argument, one disclaimed before us, that the Marleasing principle required a narrower interpretative approach than the like interpretative obligation under section 3 of the Human Rights Act 1998.
The EAT then also held that the employees’ interpretation did not go against the grain of the WTR. Langstaff J said:
‘64 … First, the Working Time Regulations 1998 were specifically made to implement the Working Time Directive. It can be presumed that the intention of Parliament was to fulfil its obligation to do so fully and accurately. If, seen through a modern lens, the words do not achieve that, then to adopt a conforming interpretation is not doing violence to the intention of Parliament but instead respecting it.’
Langstaff J acknowledged that that view produced a result contrary to that of Bamsey but he said that case pre-dated the later cases in the CJEU which decided that, contrary to the court’s understanding in Bamsey, the Directive did not leave it to the member states to decide how to calculate the remuneration payable. Langstaff J continued:
‘64. … Although Bamsey demonstrates what the interpretation of the 1998 Regulations should be if untrammelled by European Union law, it does not purport to identify a cardinal feature, guiding purpose or “grain” of the legislation which would preclude a different interpretation, such that it could confidently be said that Parliament had so set its face against that other view that it could not be adopted. …
There is nothing intrinsic to the 1998 Regulations which requires holiday pay to exclude payment for overtime which a worker has actually worked prior to holiday being taken, where the worker is contractually obliged to do the work. The essential feature of the Regulations is that holidays should be paid. That is not in issue. What is in question is, rather, the principle by which the amount of that payment is to be calculated. The obligation to construe legislation “as far as possible” to conform is a powerful one. I cannot accept that the form, nature and purpose of the Regulations makes it impossible to construe it as the claimants contend.
Though it is the effect of the interpretation, rather than the precise words which matters, a conforming interpretation is best expressed by amending regulation 16(3)(d) of the Working Time Regulations 1998 to insert the following italicised words, as the tribunal in Freightliner v. Neal [a third appeal before the EAT that was settled shortly before the hearing] thought appropriate, and as the Secretary of State for Business Innovation and Skills regards as permissible, namely: “(d) as if the references to sections 227 and 228 did not apply and, in the case of the entitlement under regulation 13, sections 223(3) and 234 do not apply.”’
The submissions
The essence of Mr Cavanagh’s submissions was that, applying the domestic rules of statutory construction, it is not possible to apply a conforming interpretation to the WTR and the provisions of the ERA they incorporate. Such an interpretation would go against their grain, be contrary to their underlying thrust and also to a cardinal or fundamental feature of the WTR. It would also run counter to fundamental principles of EU law. Whilst Mr Cavanagh disclaimed that it would require tribunals to make policy decisions for which they are not equipped (compare Ghaidan [2004] 2 AC 557, at para 33, per Lord Nicholls), he did submit that it would give rise to serious problems in relation to the practical implications of a conforming interpretation. In particular, it would raise difficulties as to the determination in any particular case of the appropriate reference period for the calculation of the employee’s ‘normal remuneration’. How, for example, does a tribunal deal with a case where the results-based commission is only payable after a defined threshold of return is achieved, which may mean that for some months of the year the employee receives no commission? How does it deal with the case of the banker who is paid a salary and a large results-based annual bonus? In addition, Mr Cavanagh said that for a conforming interpretation to be imposed upon the WTR would offend the EU principle of legal certainty under which a person should be able to work out from the domestic legislation what he needs to do in order to comply with the law; and that it would also offend the EU principle of non-retroactivity.
Mr Cavanagh opened his oral submissions by referring us to Evans and said it showed that it is not possible to apply to the WTR and the provisions of the ERA they incorporate an interpretation conforming with the ‘normal remuneration’ requirements of article 7. In enacting the WTR, Parliament had, he said, made a deliberate choice as to the definition of ‘pay’ it was adopting and it went against the grain of the WTR to re-write the provisions applying to Mr Lock’s case in the way that the ET did. Evans shows that the correct interpretation of section 221(2) is a barrier to a conforming interpretation.
Mr Cavanagh referred us to several European decisions which he said show that the principles of conforming interpretation do not justify an interpretation of domestic law that is contra legem, meaning contrary to the clear meaning of the domestic language in issue. In Centrosteel SrL v. Adipol GmbH (Case C-456/98) [2000] ECR 1-6017, Advocate General Jacobs said, in para 32 of his Opinion, that ‘… the national court is not required to interpret national law in a way which is contrary to the express terms of the relevant legislation.’ The like point was made in Adeneler and others v. Ellinikos Organismos Galaktos (Case C-212/04) [2006] IRLR 716, in which the CJEU also, however, emphasised the national court’s obligation to do all it can to adopt a conforming interpretation. The court said:
‘109. The requirement for national law to be interpreted in conformity with Community law is inherent in the system of the Treaty, since it permits national courts, for the matters within their jurisdiction, to ensure the full effectiveness of Community law when they determine the disputes before them (see, inter alia, Pfeiffer and others, paragraph 114).
It is true that the obligation on a national court to refer to the content of a Directive when interpreting and applying the relevant rules of domestic law is limited by general principles of law, particularly those of legal certainty and non-retroactivity, and that obligation cannot serve as the basis for an interpretation of national law contra legem (see, by analogy, case C-105/03 Pupino [2005] ECR I-5285, paragraphs 44 and 47).
Nevertheless, the principle that national law must be interpreted in conformity with Community law requires national courts to do whatever lies within their jurisdiction, taking the whole body of domestic law into consideration and applying the interpretative methods recognised by domestic law, with a view to ensuring that the Directive in question is fully effective and achieving an outcome consistent with the objective pursued by it (see Pfeiffer and others, paragraphs 115, 116, 118 and 119).’
The CJEU said much the same in Impact v. Minister for Agriculture and Food and others (Case C-268/06) [2008] IRLR 552, at paras 100 and 101; and said it again in Dominguez v. Centre Informatique due Centre Ouest Atlantique (Case C-282/10); [2012] IRLR 321, at paras 25 and 27. By way of a more recent reference to the ‘contra legem’ principle, Mr Cavanagh referred us to the Opinion of Advocate General the Opinion of Advocate General Bot in Dansk Industri, acting on behalf of Ajos A/S v. Estate of Karsten Eigil Rasmussen [2016]IRLR 552, [2016] IDS Pensions Law Reports 1:
‘68. The Latin expression “contra legem” literally means “against the law”. A contra legem interpretation must, to my mind, be understood as being an interpretation that contradicts the very wording of the national provision at issue. In other words, a national court is confronted by the obstacle of contra legem interpretation when the clear, unequivocal wording of a provision of national law appears to be irreconcilable with the wording of a directive. The Court has acknowledged that contra legem interpretation represents a limit on the obligation of consistent interpretation, since it cannot require national courts to exercise their interpretative competence to such a point that they substitute for the legislative authority. …
72. An obstacle that is presented by the case-law is thus not comparable to an obstacle consisting in the existence of a provision of national law whose very wording is irreconcilable with a rule of EU law. In the latter case, the obstacle cannot be overcome by the national courts without their substituting themselves for the legislative authority and re-writing the provision in question.
73. I would add that to recognise the existence of a national case-law that is contrary to EU law as an obstacle to the interpretation by national courts of a provision of national law in conformity with EU law would greatly diminish the potential of the technique of consistent interpretation in resolving conflicts between EU law and national law.’
For further guidance as to the interpretative principles, Mr Cavanagh referred in his skeleton argument to the CJEU’s decision in Pfeiffer and others v. Deutsches Rotes Kreuz, Kreisverband Waldshut eV (Cases C-397–403/01) [2005] ICR 1307, which was in particular relied upon by Mr Ford in argument for what he said is ‘the Pfeiffer presumption’ and for a clear statement that the Marleasing conforming interpretative obligation must be performed in accordance with the principles of the national law. The court said:
‘111. It is the responsibility of the national courts in particular to provide the legal protection which individuals derive from the rules of Community law and to ensure that those rules are fully effective.
That is a fortiori the case when the national court is seised of a dispute concerning the application of domestic provisions which, as here, have been specifically enacted for the purpose of transposing a Directive intended to confer rights on individuals. The national court must, in the light of the third paragraph of article 249 EC [now article 288 TFEU], presume that the member state, following its exercise of discretion afforded to it under that provision, had the intention of fulfilling entirely the obligations arising from the Directive concerned: see Wagner v. Fondo de garantia salarial (Case C-334/92) ECR I-6911, 6932, para 20.
Thus, when it applies domestic law, and in particular legislative provisions specifically adopted for the purpose of implementing the requirements of a Directive, the national court is bound to interpret national law, so far as possible, in the light of the wording and the purpose of the Directive concerned in order to achieve the result sought by the Directive and consequently comply with the third paragraph of [article 288].
Although the principle that national law must be interpreted in conformity with Community law concerns chiefly domestic provisions enacted in order to implement the Directive in question, it does not entail an interpretation merely of those provisions but requires the national court to consider national law as a whole in order to assess to what extent it may be applied so as not to produce a result contrary to that sought by the Directive: see, to that effect, Carbonari [1999] ECR I-1103, 1134–1135, paras 49 and 50.
The requirement for national law to be interpreted in conformity with Community law is inherent in the system of the Treaty, since it permits the national court, for the matters within its jurisdiction, to ensure the full effectiveness of Community law when it determines the dispute before it: see, to that effect, Mau v. Bundesanstalt fur Arbeit (Case C-160/01) [2003] ECR I-4791, 4829, para 34. …
In that context, if the application of interpretative methods recognised by national law enables, in certain circumstances, a provision of domestic law to be construed in such a way as to avoid conflict with another rule of domestic law or the scope of that provision to be restricted to that end by applying it only in so far as it is compatible with the rule concerned, the national court is bound to use those methods in order to achieve the result sought by the Directive.’
Mr Cavanagh submitted that the UK case law on conforming interpretation is consistent with the European decisions and he referred us to the key domestic authorities. He emphasised the cautionary guidance in Ghaidan and in Vodafone 2 against, in effect, the amendment of domestic legislation by way of purported interpretation. The essence of his submissions was that whatever the limits of the Marleasing interpretative obligation, it does not empower the court to repeal domestic legislation and to substitute a new provision in its place. That is, in Lord Bingham’s phrase, judicial vandalism.
He said that in this case the ET’s interpretative exercise went too far: the effect of the additional words it read into the WTR was simply, and expressly, to reverse the effect of section 221(2) of the ERA. He referred us to the decision of the Divisional Court in Regina v. British Coal Corporation, Ex parte Vardy and Others [1993] ICR 720, in which, at 751 to 753, Glidewell LJ, with whom Hidden J agreed, explained why he felt unable to interpret section 188(1) of the Trade Union and Labour Relations (Consolidation) Act 1992 in conformity with article 2 of Directive 75/129/EEC, for reasons which can perhaps be regarded as contra legem reasons, although they were not so described. I might add, however, that in the EAT’s later decision in UK Coal Mining Ltd v. National Union of Mineworkers (Northumberland Area) and another [2008] ICR 163, at para 85, Elias J, when delivering the judgment of the EAT, expressed his reservations as to the correctness of that assessment in Vardy, but noted it was made in an area that is ‘very much a matter of impression’ and said it would be wrong for the EAT to depart from it.
Mr Cavanagh said that the circumstances in which the courts have been prepared to adopt a Marleasing conforming interpretation have been quite different in character from those of the present case. They have been cases in which the interpretation has truly gone ‘with the grain’ of the legislation and has, therefore, not involved its repeal or reversal but has supplemented it. He said they fell into three categories.
First, there were cases when an examination of the domestic statute revealed a lacuna, with the result that it failed to incorporate a requirement of the Directive. He referred by way of example to the EAT’s decision in EBR Attridge LLP (formerly Attridge Law) and another v. Coleman [2010] ICR 242. The question there was whether the provisions of the Disability Discrimination Act 1995, as amended in 2004 so as to transpose the effect of the EU’s Framework Directive (2000/78/EC), could be interpreted so as cover the case of so-called ‘associative discrimination’, to which the Directive was not expressly directed but to which the CJEU held in Coleman v. Attridge Law (Case C-303/06) [2008] ICR 1128 it was in fact directed. The EAT, by Underhill J, the President, upheld the ET’s decision that the provisions in the amended 1995 Act could be read as extending to associative discrimination. The language that was necessarily written in to achieve that end did not involve any repeal or reversal of the Act’s existing language. It did no more than effect an extension of it in line with the manifest grain of the legislation. Mr Cavanagh said that this court’s decision in NHS Leeds v. Larner [2012] EWCA Civ 1034; [2012] ICR 1389 was another example of his first category.
Mr Cavanagh’s second category comprised cases in which he said the literal interpretation of the domestic language would fly in the face of the intended purpose of giving effect to the Directive and in which words were read in to make the intended conforming meaning clear. Examples were said to be the decisions in the House of Lords in Pickstone and Others v. Freemans plc [1989] 1 AC 66; and Litster and Others v. Forth Dry Dock & Engineering Co Ltd (in receivership) and Another [1989] ICR 341.
In Pickstone, the House of Lords was concerned with the interpretation of Regulation 2(1) of the Equal Pay (Amendment) Regulations 1983, which amended section 1(2) of the Equal Pay Act 1970 so as to correct the defect in its equal pay legislation identified by the ECJ in Commission of the European Communities v. United Kingdom of Great Britain and Northern Ireland (Case 61/81) [1982] ICR 578. The House held that it was entitled to read some qualifying words into Regulation 2(1) since to do so would give clear effect to the intention of the amendment as explained in Parliament, whereas their literal interpretation would mean that the Regulations, although purporting to give effect to the UK’s obligations under article 119 of the EEC Treaty, were in fact in breach of them. The justification for so implying words into the Regulations is well explained in the speech of Lord Oliver of Aylmerton, at [1989] AC 66, at 126 to 128. As Lord Rodger explained in Ghaidan, the House in Pickstone was doing no more than legitimately making an implication that went with the grain of the legislation. Lord Rodger described the House’s decision in Litster in the like way.
Mr Cavanagh’s third category was where there was a drafting error in the domestic legislation. As an example, he cited Rowstock Ltd and another v. Jessemey [2014] EWCA Civ 185; [2014] ICR 550. That raised the question whether post-employment victimisation was unlawful under the Equality Act 2010. The background was that, whereas by the time the 2010 Act was enacted it was well-established that post-employment discrimination, including victimisation, was unlawful, section 108(7) of the 2010 Act was apparently drafted in terms expressly excluding claims for such victimisation. The Act was intended to give effect to a number of EU Directives and the EAT had recognised what it called the ‘flexible interpretative approach’ required when construing legislation intended to implement EU law. Nevertheless, it concluded that it was not possible to interpret the Act as prohibiting post-employment victimisation since it considered that section 108(7) had provided in terms that such victimisation was not unlawful. The case therefore fell on the wrong side of the Ghaidan line that marked the limit of when a conforming interpretation is permissible.
This court (Maurice Kay, Ryder and Underhill LJJ) allowed the appeal against the EAT’s decision. Underhill LJ gave the leading judgment, with which Ryder and Maurice Kay LJJ agreed, and said that, whilst the natural meaning of the language of the 2010 Act was that it did not proscribe post-termination victimisation, it was equally clear that that was not the result its draftsman intended. Underhill LJ gave five reasons why the apparent failure of the Act to proscribe post-termination victimisation was ‘a drafting error’, although he said it was not possible to ascertain how it had arisen.
The question, therefore, was how far the court could go in putting the matter right. Underhill LJ referred to the interpretative approach explained in Pickstone, Litster and Ghaidan and cited the summary of the principles in Vodafone 2. He concluded, at para 42, that given the existence of the EU obligation to proscribe post-employment victimisation, the only question was whether it was ‘possible’, in the Ghaidan sense, to imply words into the 2010 Act to achieve that result. In his view, it was. The implication of such a prohibition would not only be consistent with the fundamental principles of the Act and ‘go with its grain’, it would represent what the draftsman intended.
Mr Cavanagh submitted that the present case falls outside his three categories and is of a quite different character. It is not enough to say, as was said by Langstaff J in Bear Scotland, that the WTR were intended to give full effect to the Directive. Parliament chose to implement the Directive in a specific and deliberate way that incorporated a statutory code for identifying the measure of holiday pay, a code providing unambiguously that an employee such as Mr Lock receives only basic pay by way of holiday pay. To repeal and replace that provision by way of purported conforming interpretation is to ignore the contra legem principle and to cross the line between judicial interpretation and legislation. The ‘grain’ of the WTR is to be found in the fact that regulation 16(2) incorporates a package of pre-existing rules for the determination of ‘pay’. The present case throws up a set of facts for which the Directive provides one answer and the legislation another. He said the logic of the contrary argument is that any conforming interpretation must be to the effect that sections 221 to 224 of the ERA shall apply except where they do not apply. He said the relevant grain is the adoption lock, stock and barrel of a pre-existing set of rules.
Mr Cavanagh placed considerable reliance on the decision of the House of Lords in Duke v. GEC Reliance Ltd (formerly Reliance Systems Ltd) [1988] ICR 339. I do not regard that authority as of relevant assistance for present purposes, but given the importance Mr Cavanagh attached to it I should explain why. It should be noted first that the decision pre-dated the Marleasing decision, which made clear that the conforming interpretative obligation applies not only to legislative provisions directed at implementing a Directive, but also to those adopted before the enactment of the Directive.
In Duke, the employer’s policy was to retire its women employees at 60 and its men at 65. When they retired Mrs Duke at 60, she asserted she was the victim of discrimination under section 6(2) of the Sex Discrimination Act 1975 by virtue of her dismissal. The employer’s response was that its actions were saved by section 6(4), which provided that section 6(2) did not ‘apply to provision in relation to death and retirement.’ Mrs Duke argued that section 6(4) must be construed in a sense favourable to her so as to comply with Community law. The Sex Discrimination Act 1975 and the Equal Pay Act 1970 both came into force on 29 December 1975. Lord Templeman, with whose speech all their lordships concurred, said, at 346:
‘The Government and Parliament of the United Kingdom must have considered that [both Acts] complied with the obligation of the United Kingdom to observe Community law and Community intentions including article 119 and the Equal Pay Directive, so far as Community law was understood in the United Kingdom and so far as Community intentions were discernible.’
On 9 February 1976, following the coming into force of both Acts, the Equal Treatment Directive came into force. That barred discrimination between the sexes of the type from which section 6(4) of the 1975 Act exempted the employers. The time for complying with that Directive expired in August 1978, by which time the UK had taken no steps to repeal or amend section 6(4). Mrs Duke’s argument was, however, that the two Acts of 1975 (which came into force on 12 November 1975) had to be construed in a manner giving effect to the later Equal Treatment Directive of 9 February 1976 as, much later, interpreted by the ECJ in Marshall v. Southampton and South West Hampshire Area Health Authority (Teaching) [1986] ICR 335, a decision published on 20 February 1986. That case decided that:
‘… article 5(1) of [the Equal Treatment Directive] must be interpreted as meaning that a general policy concerning dismissal involving the dismissal of a woman solely because she has attained the qualifying age for a state pension, which age is different under national legislation for men and for women, constitutes discrimination on grounds of sex, contrary to that Directive.’
As to whether, as Mrs Duke claimed, section 6(4) of the Sex Discrimination Act 1975 should be construed in a manner giving effect to the 1976 Directive as so construed in Marshall, Lord Templeman gave the argument short shrift. He said, [1988] ICR 339, at 351 to 352:
‘… The Acts [of 1970 and 1975] were not passed to give effect to the Equal Treatment Directive and were intended to preserve discriminatory retirement ages. Proposals for the Equal Treatment Directive dated 9 February 1976 were in circulation when the Bill for the Sex Discrimination Act 1975 was under discussion but it does not appear that these proposals were understood by the British Government or the Parliament of the United Kingdom to involve the prohibition of differential retirement ages linked to differential pensionable ages. …
… [the full argument before this House] has satisfied me that the Sex Discrimination Act 1975 was not intended to give effect to the Equal Treatment Directive as subsequently construed in the Marshall case [1986] ICR 335 and that the words of section 6(4) are not reasonably capable of being limited to the meaning ascribed to them by the appellant. Section 2(4) of the European Communities Act 1972 does not in my opinion enable or constrain a British court to distort the meaning of a British statute in order to enforce against an individual a Community directive which has no direct effect between individuals. Section 2(4) applies only where Community provisions are directly applicable.’
It is unnecessary to refer further to Duke. Its circumstances were fundamentally different from those of this case. It was a case in which the relevant statute had not been directed at giving effect to a Directive that was subsequently promulgated: see Lord Templeman’s own summary in his speech in the Pickstone case at [1989] 1 AC 66, at 123C. By contrast, in the present case, the WTR were plainly directed at transposing the existing Working Time Directive, later consolidated into the Directive. The teaching of Duke therefore sheds no light on the issue before us, namely whether the WTR can be interpreted conformably with the Directive, as later interpreted by the CJEU, that it unquestionably purported to transpose. There may be reasons why they cannot, but Duke does not demonstrate them.
Mr Cavanagh criticised the decision in Bear Scotland. He said the EAT was wrong, in para 64, to say that Bamsey had not identified ‘a cardinal feature, guiding purpose or “grain” of the legislation’ whose sense differed from the interpretation advanced by the appellants. Bamsey showed that the ERA legislation incorporated by the WTR made expressly clear in section 234 that non-guaranteed overtime was not to be taken into account when calculating a week’s pay; and para 40 of Bamsey shows that the grain of the ERA provisions was not consistent with matching holiday pay to normal pay. He criticised para 66 of Bear Scotland: it appeared to overlook that the WTR specifically incorporated provisions of the ERA that expressly excluded the inclusion of non-guaranteed overtime in the calculation of holiday pay. As for the assertion in para 64 that it is to be presumed that Parliament intended by the WTR to fulfil its implementation obligations fully and accurately, so enabling the courts to repair any shortcomings that it identifies through ‘a modern lens’, that went too far. The question of whether a conforming interpretation should be applied only ever arises when there is an apparent mismatch between a Directive and the implementing legislation. If the consequence is that, regardless of its language, all implementing legislation is to be interpreted in line with the Directive, there would be no limits to the operation of the conforming interpretation obligation. None of the leading authorities on conforming interpretation supported such an approach. They recognise that Parliament may have adopted a legislative choice that is clearly at odds with what is required by the Directive.
Finally, Mr Cavanagh submitted that para 40 of Bamsey included a second ratio for the decision: namely, that if there was any inconsistency between the WTR and the Directive, a conforming interpretation was not possible, since the court’s holding was that the WTR, when read with the ERA, ‘showed that their draftsman clearly intended to apply the [ERA’s] well established domestic definition of “a week’s pay” save in the immaterial respects for which he specifically provided in regulation 16(3).’ He said that in para 40 Auld LJ was, amongst other things, holding that there was nothing in article 7 that required holiday pay to be broadly equivalent to (in short) the employee’s normal remuneration (including therefore an overtime element) but that ‘in any event, sections 221 to 224, with or without section 234, will not necessarily achieve that’. That reflected a decision of the court that even if article 7 did bear the claimed meaning, the relevant provisions of the ERA could not be read conformably with it.
In his submissions for Mr Lock, Mr Ford said there was no dispute that: (i) article 7, as interpreted by the CJEU, required Mr Lock’s holiday pay to include a commission element; (ii) for the purposes of applying the Marleasing interpretative obligation, the court must apply the domestic rules of interpretation; or (iii) that those rules are as summarised in Vodafone 2, the summary in which was endorsed by the Supreme Court in Swift and in Nolan.
Mr Ford submitted that the scheme in the ERA incorporated by regulation 16 of the WTR is in fact anything but clear, but it is strictly only the WTR whose interpretation is in question, those being implementing regulations. Whilst it is said by British Gas that it is clear that the effect of section 221(2) of the ERA is to exclude an element referable to results-based commission from the calculation of an employee’s holiday pay, the first case that decided that was Evans, which post-dated the WTR by four years. As to whether an interpretation of the WTR entitling Mr Lock to a commission element in his holiday pay ‘goes with the grain of the WTR’, Mr Ford said there was only one answer, namely yes. The relevant interpretative principle permits a departure from the strict and literal interpretation of the domestic statute; it permits the implication of words necessary to enable an interpretation conforming with EU law; and the precise form of words does not matter provided only the end result is consistent with the principles summarised in Vodafone 2. He referred to Attridge and also to NHS Leeds v. Larner [2012] EWCA Civ 1034; [2012] ICR 1389, at paras 89 to 91, as illustrations of the extent of the conforming interpretative obligation.
As to the identification of the ‘grain’ of the WTR, Mr Ford said, first, that there was no dispute that they were enacted for the purpose of wholly implementing the Directive, and for no other reason. Second, the Pfeiffer presumption (in para 112 of the CJEU’s judgment) is that the national court must presume that the member state had the intention of fulfilling entirely the obligations arising under the Directive. There is nothing in this case pointing the other way. Third, the grain of the WTR is clear: the purpose of regulation 13 is for a worker to be entitled to four weeks’ annual leave in each leave year and regulation 16 is directed at ensuring that he is paid during such leave in accordance with the Directive. The adoption of the ERA scheme was a convenient scheme for complying with the Directive, and the thrust of its provisions was that leave to was to be paid. The scheme adopted a formula that had never previously been held to exclude commission. Consistent with the intention that pay should be sufficient, the WTR excluded the statutory cap that sections 227 and 228 would apply in other contexts, for example, redundancy pay or unfair dismissal.
Mr Ford submitted that in many, probably most, cases the statutory formula produces the result required by article 7. The most obvious example is the worker who is simply entitled to basic pay: he will receive the same by way of holiday pay. In many, or most, cases contractual supplements to basic pay will also be included. So will commission, which falls within the ERA’s definition of remuneration. If a worker has no normal working hours, commission is included in the averaging process. If he is a shift worker within section 222, commission is also included in the averaging. If he is a piece worker, with his commission varying with the amount of work, commission is included under section 221(3).
The only circumstance when commission is excluded is the anomalous case applying to Mr Lock, namely that of a worker falling within section 221(2). Mr Ford could not explain what, if any, policy reason there was for such a case. However, such a worker still gets holiday pay measured by reference to his basic pay. All that is being sought for him is a supplement by way of a commission element. Mr Ford submitted that the terms of section 221(2) cannot be regarded as expressing a clear intention to exclude commission. Whilst he recognised what this court decided in Evans, he pointed out that the EAT in that case had found in favour of Mr Evans by, as Mr Ford put it, adopting the approach that section 221(4) was the coach that drove section 221(3). He said Marleasing only requires the identification of a possible interpretation that conforms with EU law.
As for Mr Cavanagh’s reliance on the contra legem principle, Mr Ford said it is clear from the CJEU decisions that the principle is one recognised by the CJEU as potentially standing in the way of the adoption by the national court of a conforming interpretation. But the overriding principle is that it is for the national court to do all that is possible to achieve a such an interpretation; and it will in particular be no answer that there is established case law in its domestic law that has consistently adopted an interpretation that is incompatible with EU law. It will in such a case be necessary for the national court to change its case law. In that context, Mr Ford referred us to paras 31 to 34 of the judgment of CJEU in the Dansk case, [2016] IRLR 552, which made that clear. Since it is also clear that it is a matter for the determination of the national court whether a conforming interpretation is barred by the contra legem principle, there is nothing in the EU cases that is inconsistent with the broad approach to conforming interpretation summarised in Vodafone 2. The cases referred to by Mr Cavanagh show that the most that the CJEU ever does is to question whether the national court has taken too narrow an approach towards its interpretative obligation.
Mr Ford referred us to Revenue and Customs Commissioners v. IDT Card Services Ireland Ltd [2006] EWCA Civ 29; [2006] STC 1252, in which he said Arden LJ’s observations, at para 113, reflect something of a parallel with the circumstances of this case. She said:
‘This is not beyond the bounds of permissible interpretation because there is no indication that Parliament specifically intended to depart from the Sixth Directive in this respect. The provisions of Sch 10A are equally consistent with Parliament not having foreseen the particular problem that has arisen in this case. It follows from the Ghaidan case that the court’s duty arises even if Mr Lasok is correct in submitting that the correct interpretation of Sch 10A is that VAT is not imposed on the United Kingdom distributors of ICSIL’s phonecards in the circumstances of this case. It also arises even if Parliament did not intend to limit relief in any way for which Customs and Excise now contend. The provisions of Sch 10A do not contain any fundamental feature inconsistent with reading into para 3(3) a further disapplication of the disregard in para 3(2) to make para 3 conform to the objectives of the Sixth Directive: it is simply a case of widening the existing provision in para 3(3) …’.
Mr Ford said that provided a close analogue to the problem raised in this case by the WTR. Here too there is no indication that Parliament ever intended to depart from the Directive.
Mr Ford disagreed with Mr Cavanagh that there is a rigid threefold category of cases into one of which a case must fall before it is capable of being the subject of a conforming interpretation. He said the question in every case is simply whether the principles summarised in Vodafone 2 justify the adoption of a conforming interpretation. He submitted that the present case is not relevantly distinguishable from Bear Scotland and that the EAT’s reasoning in that case should be regarded as applying similarly to this one. The WTR must be presumed to have intended to give effect to the Directive, and in most cases the WTR do achieve the result required by the Directive of ensuring that workers receive their normal remuneration on leave. Fulfilment of the Directive is the ‘underlying thrust’ of the WTR; there is no cardinal feature of the WTR suggesting that Parliament had set its face against a conforming interpretation; and to apply a conforming interpretation is in line with the principles identified in Pfeiffer and Vodafone 2.
Finally, as to Bamsey, Mr Ford disputed that it binds this court to hold that a conforming interpretation is not possible. The court in Bamsey wrongly decided that article 7 laid down no requirements as to the nature of the level of an employee’s holiday pay entitlement. It was for that, erroneous, reason that Auld LJ said in para 40 that there was ‘nothing in regulation 16 on which the Marleasing … principle of construction can bite …’. Bamsey did not decide that a conforming interpretation was not possible. Auld LJ’s view was that the Directive left the determination of the level of an employee’s holiday pay exclusively to the national court so that there was nothing in the Directive requiring the national court to subject the WTR to a conforming interpretation examination. As the CJEU’s later decisions showed, Bamsey’s premise was wrong and so, therefore, was its conclusion. That cannot provide the basis for a decision binding this court to hold that a conforming interpretation is not possible. Langstaff J had correctly decided in Bear Scotland that Bamsey provided no authority binding upon the EAT.
Mr Tolley, for the Secretary of State, adopted Mr Ford’s submissions. He too said that a conforming interpretation is both possible and correct in this case. He answered British Gas’s argument that the adoption of a conforming interpretation would offend the principles of non-retroactivity and legal certainty by citing Sir Andrew Morritt C’s observations in Vodafone 2:
‘56. … First, it is inevitable that a conforming interpretation will be retrospective in its operation. Unless and until it is averred that the legislation is inconsistent with some enforceable Community right there is no occasion to consider a conforming interpretation. The fact that the effect of such an interpretation is felt retrospectively is no more an objection in the field of conforming interpretation than it is in the case of domestic statutory construction.
Second, it is not a requirement of a conforming interpretation that it should be capable of precise formulation. That is precisely the point summarised in sub-paragraph (f) quoted in para 37 above [see para 38 above in this judgment]. The dicta there referred to were made in such widely diverse situations as equal pay, right to succession of a protected tenancy and the imposition of a liability to VAT. It is inevitable that the conforming interpretation will lack the crispness to be expected of properly considered legislation; but, that cannot be a sufficient objection.’
Mr Tolley submitted that the decision in Pfeiffer contains the clearest modern statement from the CJEU on the application of conforming interpretation principles, which he said was the same as that explained in relation to section 3 of the Human Rights Act 1998 in Ghaidan. Pfeiffer shows that it is presumed that the member state intended by its transposing legislation to fulfil entirely its obligations under the Directive, although Mr Tolley disclaimed any suggestion that this means that a conforming interpretation will be possible in every case. He referred to the EAT’s decision in Innospec Ltd and others v. Walker [2014] ICR 645 where Langstaff J, at para 53, said that ‘what is necessary is seeing whether Parliament clearly adopted a legislative choice which was directly at odds with that required by the Directive.’ Mr Tolley said this is a matter of objective inference. Putting it another way, he said the court may ask whether there is any indication that Parliament specifically intended to depart from the Directive, or whether the domestic provisions are equally consistent with Parliament not having foreseen the particular problem in question: for that approach, he too invoked the observations of Arden LJ in the IDT Card Services case [2006] EWCA Civ 29; [2006] STC 1252, at paras 112 and 113.
Applying the guidance to be derived from the authorities to which I have referred, Mr Tolley said there is nothing intrinsic to the WTR requiring holiday pay to exclude payment reflecting variable commission that is directly linked to results, and nothing to prevent an interpretation of the WTR that enables the ‘normal remuneration’ principle to be applied in a case such as Mr Lock’s. There is no basis for attributing to Parliament a legislative choice directly at odds with what is required by article 7 of the Directive. He said the timing was important. The WTR came into force on 1 October 1998, in time to comply with the transposing obligations under the Working Time Directive. Article 7 says nothing express as to the requirement of ‘normal remuneration’. Robinson-Steele, decided by the CJEU on 16 March 2006, was the first decision to refer (in para 50, but without full explanation) to the ‘normal remuneration’ requirement. The CJEU’s decision of 20 January 2009 in Stringer and others v. Revenue and Customs Comrs (Case C-520/06) [2009] ICR 932 made a like reference, at para 58, but added no more explanation. It was only in British Airways plc v. Williams and others (Case C-155/10) [2012] ICR 847, a decision of 15 September 2012, that the CJEU first explained in detail the meaning of ‘normal remuneration’, identifying the components of the pay that ought to be included in it (see paras 21 to 27). In these circumstances, Mr Tolley submitted that there is here no basis for an inference that, in enacting the link between the WTR and the ERA, Parliament specifically intended to depart from the Directive. A more likely inference is that no-one foresaw the future explanations from the CJEU as to the meaning of ‘pay’ in article 7.
Discussion and conclusion
Whilst Mr Cavanagh dealt with it at the end of his submissions, I shall deal first with his submission that this court in Bamsey made a decision binding upon us that the WTR are not capable of interpretation in a manner conforming with article 7 of the Directive. If Mr Cavanagh is right, the appeal must succeed.
The EAT in Bear Scotland did not regard itself as bound by Bamsey in such respect, nor did either tribunal below in this case. Bamsey was, as I have said, decided before the first CJEU case explaining that concealed within the language of article 7 is a requirement binding on member states to ensure that holiday pay is calculated by reference to a worker’s ‘normal remuneration’. Given that circumstance, it is unsurprising that the court in Bamsey proceeded on the basis that the Directive, by the language of article 7, delegated to the member state all matters relating to the provision and calculation of holiday pay and that there was nothing in the Directive requiring the WTR to be interpreted otherwise than through a domestic lens. Auld LJ explained this in paras 34 to 42 of his judgment, under the heading ‘Construction of regulation 16 in the light of the Working Time Directive’. It is, however, said that in para 40 he identified two ratios for the court’s decision, of which one was that if, contrary to his view, article 7 did impose an EU-wide ‘normal remuneration’ requirement, the WTR could anyway not be read conformably with it.
In making that submission, Mr Cavanagh engaged in a close analysis of the four sentences of para 40, suggesting en route that the true sense of the ‘especially’ in the first sentence was ‘in any event’, a suggestion I would not be prepared to adopt. I can detect nothing in Auld LJ’s first sentence that can be read as deciding that, if the Directive means something different from the court’s understanding, an interpretation of the WTR conforming with it was not possible. The second sentence repeats Auld LJ’s conclusion that article 7 has no ‘normal remuneration’ requirement. With respect to Auld LJ, I am not entirely clear as to the full intended sense of his third sentence, namely ‘And, in any event, sections 221 to 224, with or without section 234, will not necessarily achieve that.’ I would, however, decline to conclude that included within it is a decision that an interpretation of the WTR conforming with a ‘normal remuneration’ requirement is not possible. The fourth sentence adds nothing material.
In my judgment, there is no basis on which it can be divined from the decision in Bamsey that the court was deciding that it is not possible to interpret the WTR conformably with any (if any there was, which the court did not accept) ‘normal remuneration’ requirement of article 7. Had it been the court’s intention so to decide, it would have said so positively and clearly. It did not do so, for the reason that it was not so deciding. It decided no more than that article 7 was irrelevant to the interpretative exercise before it, which was exclusively one of domestic statutory interpretation. I would reject Mr Cavanagh’s submission that Bamsey binds us to allow the appeal. In fairness to Mr Cavanagh, he advanced the submission with express moderation.
Turning to Evans, I have said that I am satisfied that it shows that, if the WTR and the provisions of the ERA they incorporate are interpreted exclusively through a domestic lens, Mr Lock is not entitled to holiday pay calculated by reference to his ‘normal remuneration’, that is by reference also to the commission he usually earns. Evans also pre-dated the first of the CJEU decisions revealing the true sense of article 7; and the question now before us is whether it is possible to interpret the WTR in a way that conforms with the article 7 requirement as so revealed. There is no dispute that Evans does not bind us to answer that question in the negative. One element of the contra legem principle referred to by the EU decisions requires us to leave Evans out of account when approaching the task of finding, if possible, a conforming interpretation: see the judgment in Dansk, para 33.
Another element of the contra legem principle reflects the acceptance by the CJEU that, when engaging in a conforming interpretation exercise, a national court may find it impossible to adopt a conforming interpretation when ‘the clear unequivocal wording of a provision of national law appears to be irreconcilable with the wording of a directive’ (see the Opinion of Advocate General Bot in Dansk, para 68). Mr Cavanagh relied on that in support of his submission that this court is faced by just such a problem. This is a case, he says, in which the WTR adopted a statutory scheme in the shape of the incorporated provisions of the ERA; and the plain meaning of the provision relevant to this case, namely section 221(2), presents an insurmountable barrier to an interpretation of the WTR conforming with article 7.
As it seems to me, if this element of the contra legem principle as explained by the EU cases is applied at anything approaching face value, it would be likely to frustrate the possibility of a conforming interpretation in many cases. For example, I find it difficult to see how the House of Lords could have decided Pickstone as it did had it applied the contra legem principle as so explained. Lord Oliver explained how it was not possible, according to ordinary domestic canons of construction, to interpret the words in question in a conforming way, but he also explained why it was nevertheless appropriate so to interpret them as appropriately modified (see [1989] 1 AC 66, at 126). Pickstone was later rationalised by Lord Rodger in Ghaidan [2004] 2 AC 557, at para 121, as one in which the House had interpreted the critical words in line with the grain of the legislation. In the same case, at paras 29 to 33, Lord Nicholls explained how a conforming interpretation for the purposes of section 3 of the Human Rights Act 1998 may require a departure from the unambiguous meaning the statute might otherwise bear, whilst also explaining how the wide scope for a conforming interpretation is circumscribed by the limits of the grain or underlying thrust of the legislation.
What emerges from Ghaidan and the summary in Vodafone 2, the latter having since been endorsed by the Supreme Court in Swift and Nolan, is that the United Kingdom has dealt with the contra legem principle in a manner that is manifestly more in line with the EU objective of conforming interpretation at member state level than might be the case by anything approaching a rigid application of the principle summarised by Advocate General Bot in Dansk. When faced with the question of whether a conforming interpretation can be adopted, the courts of the United Kingdom do not confine themselves to a consideration of the literal meaning of the language that may appear to stand in their way; they approach the task by reference to the broader considerations of whether a conforming interpretation will be in line with the grain or underlying thrust of the legislation. That is an approach that ought, I would think, to attract nothing but commendation by the CJEU.
I do not, therefore, derive assistance from Mr Cavanagh’s contra legem submissions. In my view the critical question comes down to whether the conforming interpretation of the WTR for which Mr Lock contends is or is not within the grain or underlying thrust of that legislation. If it is, I consider it ought to follow that the interpretation favoured by the tribunals below is one this court should uphold. If it is not, a conforming interpretation is not possible.
I have not found this question easy and my view on the answer has wavered. I should perhaps first clarify what it is I consider the court is concerned to interpret. It is not the provisions of sections 221 to 224 of the ERA read in isolation and apart from the WTR. Rather, it is the WTR and those provisions of the ERA that they incorporate as, in effect, in a schedule.
Mr Cavanagh’s persuasive argument that a conforming interpretation is not possible merits serious consideration. The WTR pre-dated the CJEU decisions that have explained article 7’s requirement for holiday pay to match the worker’s ‘normal remuneration’. They were enacted under a Directive whose language apparently delegated to the member state the right to determine all aspects of holiday pay, including its calculation. They did so by incorporating a suite of statutory provisions relating to the calculation of ‘a week’s pay’ dating from 1963, being provisions which in at least two identified respects (cases like Mr Lock’s and the non-guaranteed overtime cases) do not provide for holiday pay to match ‘normal remuneration’. How, in those circumstances, can it be said that the grain or underlying thrust of the WTR and the incorporated provisions of the ERA is other than one directed simply at measuring a worker’s holiday entitlement by whatever those provisions provide. If they discriminate in any respect between the pay entitlements of workers who are engaged on different terms, that may be unfair; but if that is what they provide, that is the end of the matter. How can implementing provisions conceived in the circumstances just described be construed in conformity with requirements of the Directive that were only first explained by the CJEU some eight years later?
There is, however, also another way of answering the critical question, one that I have concluded that I prefer. There is no dispute that the WTR were enacted solely and deliberately for the purpose of implementing the requirements of the Directive; and I agree with Mr Ford and Mr Tolley that the Pfeiffer presumption requires the court to presume that the United Kingdom government intended by the WTR to fulfil entirely the obligations arising under the Directive. That presumption also encompasses an intention to fulfil even those requirements of the Directive which were not apparent at the time of the enactment of the Directive, but which only became clear by later elucidation by the CJEU. Since the enactment of the WTR, the CJEU has explained the true requirements of article 7, an explanation of which the United Kingdom government could not reasonably have been aware when it enacted the WTR.
In fact, however, at least in the case of most types of worker, the WTR do provide for the ‘normal remuneration’ measure (including by reference to the commission they earn) for the purpose of calculating holiday pay. So far as counsel are aware, there are just two, apparently anomalous, exceptions to that, namely in relation to (i) workers such as Mr Lock employed on terms to which section 221(2) applies, and for whom their commission payments are not taken into account in the calculation; and (ii) workers with non-guaranteed overtime. At least as to the former class, it is by no means clear how apparent the existence of this exception was when the WTR were enacted: it was not until four years later that the point was explained by Evans.
Even given the Pfeiffer presumption, I readily accept, however, that it does not automatically and necessarily follow that a conforming interpretation of implementing domestic legislation will be possible in every case. It is, I consider, still necessary to apply an objective assessment as to whether a legislative choice has been made that is directly at odds with the requirements of the Directive. Mr Ford was, I consider, correct to accept that had the WTR expressly and unambiguously provided that a worker engaged on Mr Lock’s terms was not to have commission taken into account when calculating his holiday pay, then Mr Lock would have no case.
In my view, however, that cannot be said in relation to the WTR. Rather, I would regard their provisions as more consistent with the legislature simply not having foreseen the particular problem that was in due course to arise with the subsequent decisions of the CJEU as to the true sense of article 7. I regard the case as in line with the type of circumstances that Arden LJ was considering in the IDT Card Services case, at para 113.
This takes me to the conclusion that the WTR are properly to be regarded as in the nature of implementing provisions that, in nearly all respects, properly implement the Directive as subsequently explained; but that in two anomalous types of case (Mr Lock’s case being one) provide for a lower measure of holiday pay than article 7 in fact requires. I am not prepared to conclude that these two anomalous cases reflect a positive legislative choice deliberately directed at discriminating against the two types of worker in the calculation of their holiday pay. No one has suggested any reason for such discrimination. As a matter of objective inference, I regard it as more likely that the differential treatment inherent in the scheme of the WTR was simply not foreseen at the time they were enacted.
I have therefore concluded that this is a case in which the grain or thrust of the WTR can fairly be identified as directed at providing holiday pay for workers measured by reference to criteria required by article 7 as since explained by the CJEU; and that, in line with that grain or thrust, the court can, and should, interpret the WTR as providing that Mr Lock is also entitled to have his holiday pay calculated by reference to his normal remuneration. To do so is to do no more than to interpret the WTR as also requiring his commission earnings to be taken into account when calculating his holiday pay. So to interpret the WTR does of course require the implication into it of words necessary to make that meaning clear. But so to imply such words is not a judicial exercise amounting to the repeal or amendment of the legislation. It is rather an example of the court performing its duty to provide a conforming interpretation to legislation introduced for the purpose of implementing a Directive.
I would, therefore, conclude that the ET was correct to interpret the WTR as it did and that the EAT was correct to uphold the decision. I add that I was also not persuaded by Mr Cavanagh that so to interpret the WTR will involve any infringement of the EU principles against retroactivity or of certainty. The result is that I would dismiss British Gas’s appeal.
I add this. I have quoted the words that the ET held should be read into the WTR and the reference period that it directed should apply for determining Mr Lock’s full holiday pay entitlement (see para 26 above). In the course of the argument, there was some discussion about how a conforming interpretation of the WTR might apply to different types of case. The court was, for example, exercised by the case of the salaried banker who receives a single, large results-based annual bonus in, say, March. Is he entitled on his summer holiday to leave pay including an element referable to his bonus? And how does or ought the WTR deal with the type of worker who is employed on terms like Mr Lock’s, but who only becomes entitled to commission at the point in the year when a particular level of turnover, profit or other threshold is reached, which may mean that he receives no commission for some months of each year? Other types of case will raise other questions.
My response to questions such as these – and to others covering other situations – is that nothing in this judgment is intended to answer them. It is no part of this court’s function to do more than to deal with the instant appeal. In the case of the banker example, there may indeed be a question as to what his ‘normal remuneration’ is, and whether its calculation ought to reflect the fact of his annual bonus and, if so, how. There may also be questions as to what, in any particular case, is the appropriate reference period for the calculation of the pay. I say nothing about any of that.
This judgment is, therefore, confined to Mr Lock’s case. There was also some discussion as to whether, even recognising that, the form of wording that the ET’s judgment implied into the WTR was strictly appropriate. I agree with Mr Cavanagh that the wording was expressed too widely insofar as it refers to all types of commission, and not just to contractual ‘results-based commission’ that is the subject of Mr Lock’s case. I would, therefore, favour an appropriate amendment to the ET’s judgment that will more clearly confine it to the circumstances of his case.
Disposition
Subject to those last observations, I would dismiss the appeal.
Lady Justice Gloster:
I agree.
The Master of the Rolls:
I also agree.