1. 

This case comes before the Court by way of a reference for a preliminary ruling from the Faglige Voldgiftsret, an arbitration board sitting in Copenhagen which has jurisdiction in industrial disputes concerning collective agreements. It raises questions concerning the application of the principle of equal pay for men and women laid down in Article 119 of the EEC Treaty and in Council Directive 75/117/EEC ( 1 ) (hereinafter ‘the Directive’) to a piece-work pay scheme.

2. 

This is the first time, to my knowledge, that the Court has had to address this issue.

3. 

The questions arose in proceedings between Specialarbejderforbundet i Danmark (Union of Semi-Skilled Workers in Denmark, hereinafter ‘the plaintiff’) and Dansk Industri (Confederation of Danish Industry), acting on behalf of Royal Copenhagen A/S (hereinafter ‘the defendant’), on the following issues.

4. 

Three main groups of workers are involved in the manufacture of ceramic products at Royal Copenhagen A/S: turners, painters and unskilled workers. They are all covered by the same collective agreement. Under that agreement, the workers are in principle paid on a piece-work basis, their pay being wholly or partially dependent on their output. They may however opt to be paid a fixed hourly rate which is the same for men and women and identical for all groups. The great majority of workers in the first two groups are paid at piece rates (70%): their pay consists of a fixed element which is paid as a basic hourly wage and a variable element which is paid for each piece produced. The rates of pay are fixed by negotiation between the organizations representing the workers and the undertaking in collective agreements.

5. 

This case concerns in particular: within the group of turners paid at piece rates, the sub-group of automatic-machine operators (26 persons, all men); within the group of painters paid at piece rates, the sub-group of blue-pattern painters (156 persons, of whom all but one are women); and the sub-group of ornamental-plate painters (51 persons, all women).

6. 

Let us call them, respectively, groups A, B and C.

7. 

The applicant brought proceedings before the Ligestillingsrådet (Council for Equal Treatment) of Denmark claiming that, since the average hourly piece-work pay is less for group B than for group A, the defendant has infringed the requirement of equal pay. In an opinion given on 1 May 1989, the Ligestillingsrådet found as follows: ‘The majority view within the Council is that the possibility that the law on equal treatment has been infringed cannot be ruled out’.

8. 

The case was thereupon brought before the arbitration board, which submitted the following questions to the Court:

9.

The admissibility of this reference for a preliminary ruling could be uncertain in so far as the national arbitration board's status as a court or tribunal, within the meaning of Article 177 of the EEC Treaty, may be open to question.

10.

The arbitration board was set up as part of the statutory system for the settlement of industrial disputes, in accordance with paragraph 77 of the agreement between the parties, paragraphs 10 and 22 of Danish Law No 317 of 13 June 1973 on the Labour Court, and the rules adopted on 17 August 1908 by the Danish Employees' Union and the Danish Employers' Association.

11.

In Vaassen-Göbbels, ( 2 ) the Court listed the features by which a court or tribunal within the meaning of Article 177 of the Treaty may be identified: the legal origins of its judicial authority, its permanence, the existence of an adversary procedure similar to that on which proceedings in ordinary courts of law are based, the binding nature of its decisions, and its obligation to apply rules of law.

12.

On the basis of that case-law, the Court held a Danish arbitration board of the same kind as that which made this reference to be a court or tribunal of a Member State within the meaning of Article 177: ( 3 )

‘... according to Article 22 of the Danish Law No 317 of 13 June 1973 on the Labour Court, disputes between parties to collective agreements are, in the absence of special provisions in such agreements, subject to the Agreed Standard Rules adopted by the Employers' Association and Employees' Union. An industrial arbitration board then hears the dispute at last instance. Either party may bring a case before the board irrespective of the objections of the other. The board's jurisdiction thus does not depend upon the parties' agreement.

The same provision of the aforementioned law governs the composition of the board and in particular the number of members who must be appointed by the parties and the way in which the umpire must be appointed in the absence of agreement between them. The composition of the industrial arbitration board is thus not within the parties' discretion.

In those circumstances the industrial arbitration board must be regarded as a court or tribunal of a Member State within the meaning of Article 177 of the Treaty’. ( 4 )

13.

Consequently, there is no doubt that the Faglige Voldgiftsret has standing to submit a question to the Court for a preliminary ruling.

14.

In the order for reference, the arbitration board states that the reference to the Court of Justice does not relate to the issue of the value of the work done. ( 5 )

15.

Since it is for the national court to assess the relevance of the questions submitted to the settlement of the dispute before it, ( 6 ) the Court of Justice cannot disregard the information furnished by that court in the order for reference, which delimits the factual and legal context of the questions submitted. ( 7 )

16.

The Court is thus obliged to base its ruling on the premiss stated by the arbitration board that the work of each of the groups to be compared, although different in nature, must be assumed to be of equal value.

17.

Admittedly that assumption leaves me somewhat puzzled, but I shall merely draw the attention of the arbitration board, whose responsibility it is to verify the truth of such a premiss, ( 8 ) to the fundamental importance of the fact that there must be work of equal value, within the meaning of Article 1 of the Directive, for the purposes of establishing discrimination on the grounds of sex with regard to pay. ( 9 )

18.

In the present case, discrimination would consist in the different treatment of two identical situations. Thus, before considering the possibility of discrimination, the national court must make sure that the groups to be compared are in identical situations.

19.

By its first question the arbitration board seeks to know whether the principle of equal pay for men and women, laid down by Article 119 of the Treaty and in the Directive, is applicable to remuneration at piece-rates.

20.

It cannot be denied that a pay system under which the level of pay is wholly or partially dependent on the output of each worker (piece-work) is encompassed by the definition of ‘pay’ in the first paragraph of Article 119 of the Treaty to which the Court has given a wide interpretation. ( 10 )

21.

That article lays down the principle that men and women should receive equal pay for equal work. In subparagraph (a) of the third paragraph, it states that equality in that context means ‘that pay for the same work at piece rates shall be calculated on the basis of the same unit of measurement’.

22.

Thus the very wording of Article 119 of the Treaty provides for the application of the principle to piece-work systems.

23.

As regards the Directive, which ‘in no way alters the content or scope of (the) principle’ defined by Article 119, ( 11 ) Article 1 provides that the principle of equal pay is applicable to ‘all aspects and conditions of remuneration’.

24.

Consequently, the first question must be answered in the affirmative.

25.

Once the national court has made certain that the two groups for comparison do indeed perform work of equal value (see my preliminary remarks), it still must verify — as the Commission points out in its observations ( 12 ) — that the choice and composition of the groups are not fortuitous, and that it is appropriate to compare them.

26.

If the national court's examination of those questions leads it to conclude that the composition of the two groups is so disparate that they manifestly cannot be compared, consideration of the following matters would ipso facto be rendered nugatory.

That is why I propose that the first question to be examined should be that concerning the requirements which may be laid down regarding the composition of the two groups to be compared.

27.

As regards the workforce of the defendant undertaking, the arbitration board provided the following information:

28.

According to settled case-law it is not only possible, but also necessary, to lay down requirements regarding the composition of the groups to be compared. ( 13 )

29.

That is moreover only logical since, if the comparison is to be relevant, it must involve representative groups of workers. ( 14 ) It would be unacceptable simply to endorse a choice which is deliberately designed to buttress the arguments of one of the parties.

30.

The fact remains that such an assessment is a question of fact within the exclusive competence of the national courts. It would obviously be inappropriate for the Court to fix a minimum number of workers for each group or a percentage threshold above which a group would be representative of an undertaking's total workforce.

31.

I shall simply mention that the comparison must be made ‘in relation to a relatively large number of employees’ ( 15 ) and that ‘it is for the national court to assess whether it may take into account those statistics, that is to say, whether they cover enough individuals, whether they illustrate purely fortuitous or short-term phenomena, and whether, in general, they appear to be significant’. ( 16 )

32.

In view of those requirements, it is understandable that the arbitration board raised questions concerning the evidential value of a comparison between two groups whose membership, both in terms of numbers (one group consists of 26 persons and the other of 156) and of proportionality in relation to the group to which each belongs (18% in the former case, 49% in the latter), is so different.

33.

In the present case, it may be asked whether the number of workers making up group A is not too low. On that point it will be recalled that in Danfoss ( 17 ) it was on that ground that the arbitration board rejected the applicant's submissions (the comparison lay between groups of 16 and of 11 persons).

34.

Furthermore, it will be necessary to assess the factual considerations which might support the plaintiff's separation of groups A and B from the other employees in its pleadings. Are those groups really the only ones to undertake work of a particular nature?

35.

Consequently, in my view the arbitration board will only have occasion to examine the subsequent questions if it first decides that the composition of the groups is in fact so representative as to provide a basis for meaningful comparison.

36.

By its second question, which is of fundamental importance and genuinely novel, the arbitration board asks whether, under a pay system based on piece-work, average hourly pay constitutes a suitable and relevant parameter for comparison.

37.

Let me straightaway express very strong reservations regarding that suggested basis for comparison in this particular context of remuneration at piece-rates.

38.

Although this is the first time that the Court has addressed this issue, it has already developed a considerable volume of case-law on equal pay for men and women paid at time rates.

39.

Allow me to recapitulate the main principles of that case-law before considering its application to the present case.

40.

The Court adopts a two-stage approach: the detection, on the basis of the facts provided, of possible discrimination, followed by a search for possible justifications. ( 18 )

41.

The Court draws a distinction between direct discrimination, which may be identified solely by reference to the criteria of equal work and equal pay, and indirect or disguised discrimination, which can only be identified by reference to implementing provisions of a Community or national character. ( 19 )

42.

‘Direct discrimination arises where Community or national legislation applies an unlawful differentiating criterion or applies formally identical rules to cases which are different. Indirect discrimination arises where Community or national legislation, while not applying any unlawful differentiating criteria, nevertheless applies other differentiating criteria the effects of which are identical or at the very least analogous to the effects which would have been produced by the application of an unlawful differentiating criterion, or, again, where Community or national legislation makes a purely formal distinction between cases which are different but in practice accords them identical treatment’. ( 20 )

43.

Direct discrimination may be identified solely by reference to the criteria of equal work and equal pay referred to in the first paragraph of Article 119 of the Treaty. ( 21 )

44.

Indirect discrimination, however, is more difficult to identify. That is precisely the problem in the present case since the remuneration at issue is not on the face of it discriminatory and applies equally to men and women.

45.

Until now, the question of the existence of indirect discrimination has chiefly been considered in the field of part-time work, which affects a considerably lower proportion of men than of women. ( 22 )

46.

That has led the Court to recognize that the total remuneration paid is a factor which may reveal discrimination. Thus the Court's judgment in Bilka ( 23 ) lays down the following rule: unequal treatment arises whenever the total remuneration paid to full-time workers is higher, hour for hour, than that paid to part-time workers.

47.

The Court likewise confirmed that the average remuneration paid may disclose a suspect situation: ‘where an undertaking applies a system of pay which is totally lacking in transparency, it is for the employer to prove that his practice in the matter of wages is not discriminatory, if a female worker establishes, in relation to a relatively large number of employees, that the average pay for women is less than that for men’. ( 24 )

48.

Lastly, statistics may raise a presumption of discrimination: ‘Where significant statistics disclose an appreciable difference in pay ... Article 119 of the Treaty requires the employer to show that that difference is based on objectively justified factors ...’. ( 25 )

49.

However, a situation which is thus disclosed as suspect does not automatically constitute an instance of indirect discrimination. It can only give rise to a ‘presumption — open to rebuttal — of incompatibility with overriding provisions on equality’. ( 26 )

50.

According to the case-law, that presumption is rebuttable in the sense that it is for the employer to prove that the measures adopted are ‘based on objectively justified factors unrelated to any discrimination on the grounds of sex’ ( 27 ) or can be explained ‘by factors which exclude any discrimination on the grounds of sex’. ( 28 )

51.

Determination of those objective factors requires a scrupulous examination of the facts, which it is for the national court to undertake. ( 29 ) While this Court has not limited the interests capable of justifying unequal treatment, it has identified the following criteria for the guidance of the national court: the measures chosen must correspond to a real need on the part of the undertaking, and they must be appropriate with a view to achieving the objectives pursued and necessary to that end. ( 30 )

52.

Thus the Court has ruled that the Directive does not prohibit ‘the use, in a job classification system for the purpose of determining rates of pay, of the criterion of muscle demand or muscular effort or that of the heaviness of the work if, in view of the nature of the tasks involved, the work to be performed does require the use of a certain degree of physical strength, so long as the system as a whole, by taking into account other criteria, precludes any discrimination on grounds of sex’. ( 31 )

53.

On the other hand, it is clear from the consistent case-law and from Article 4 of the Directive that collective agreements, by the same token as provisions laid down by law or national regulation, are subject to the prohibition of any sex discrimination with regard to pay. ( 32 ) Consequently, such agreements cannot be cited in justification of unequal treatment: ‘the fact that the rates of pay at issue are decided by collective bargaining processes ... does not preclude a finding of prima facie discrimination where the results of those processes show that two groups ... are treated differently’. ( 33 )

54.

Generally speaking, a difference in pay may be justified on a number of grounds: length of service, in all cases; training, if such training is of importance for the performance of the specific tasks which are entrusted to the employee; mobility, in the sense of adaptability to variable hours and varying places of work. ( 34 )

55.

In any event, if the courts are to be able to exercise effective supervision, the principle of equal pay must be observed in the aggregate in relation to every aspect of remuneration: ‘... it is not necessary for each individual criterion to be drawn up in such a manner as to rule out discrimination, but the system as a whole must be framed with that purpose in mind’. ( 35 )

56.

Lastly, I should like to point out that the Commission submitted a proposal for a directive on the burden of proof in the area of equal pay and equal treatment for men and women, which essentially repeats the principles set out in the Court's case-law. ( 36 ) Article 3(2) of the proposed directive provides for the establishment of a rebuttable presumption of discrimination on the basis of ‘a fact or a series of facts which would, if not rebutted, amount to direct or indirect discrimination’, it being for the respondent to prove that there has been no breach of the principle of equal treatment.

57.

Before considering the application of the principles set out by the Court to the case before us, it seems essential to draw a clear distinction between remuneration based on time rates — the only systems so far brought before the Court — and remuneration at piece-rates, such as in the present case.

58.

That was, moreover, a point which concerned the authors of the Treaty since, as noted in point 21 above, Article 119 of the Treaty is at pains to draw a clear distinction between the two situations. In Jenkins, cited above, the Court itself carefully delimited the scope of the principles it laid down to ‘work paid at time rates’. ( 37 )

59.

A worker who is paid on the basis of time worked will receive fixed pay for each unit of time worked.

60.

However, the pay of a worker on piecework is wholly or partially dependent on his productivity. As in the case of the employees in groups A, B and C, his pay may include a fixed portion, constituting a basic hourly wage. But that is always only part of what he receives. The remainder — the essential portion of his pay — is variable and paid by reference to the pieces produced. Thus, his actual earnings are determined, at least in part, by his output, which is assessed on a wholly individual basis only, according to the number of pieces completed per unit of time (work rate) and the absence of flaws in the pieces produced (quality of work).

61.

That clarification having been made, I shall now examine, as the national court requests, whether the ‘average hourly earnings’ may serve as a suitable basis for comparison of groups A and B. The figures provided in the order for reference are the average figures in each group for the fixed and the variable elements of their pay. ( 38 )

62.

As noted earlier, ( 39 ) the Court established in Enderby that ‘significant statistics’ may raise a presumption of discrimination. In accordance with that case-law, may average earnings be regarded as ‘significant statistics’?

63.

As we have seen, ( 40 ) that possibility was accepted by the Court in Danfoss.

However, it is a possibility envisaged by the Court only in so far as ‘an undertaking applies a system of pay which is totally lacking in transparency’. ( 41 ) The pay system at issue in Danfoss was one in which a worker ‘is unable to identify the reasons for a difference between (her) pay and that of (another person) doing the same work (and) ... employees do not know what criteria in the matter of supplements are applied to them and how they are applied’. ( 42 ) That does not seem to be true of groups A and B who know the details of their earnings. In the present case there does appear to be transparency since the rates of pay set are the result of agreements freely negotiated between the union and the undertaking concerned.

Secondly, Danfoss did not concern remuneration at piece-rates but a system of individual supplements applied to basic pay at time-rates.

64.

But, precisely, the extent to which the two pay systems differ has been shown above.

If average earnings were to be accepted as a valid basis for comparison, ultimately there would be a refusal to acknowledge the specific nature of remuneration at piece-rates, and no account would be taken of individual differences in output. I would emphasize the danger of that approach, which would place ‘piece-work’ on the same footing as ‘work at time-rates’ and might well raise a presumption that all remuneration at piece-rates is contrary to the principle of equal pay.

The purpose of introducing a pay system of that type in an undertaking is to help and encourage each worker to earn more by producing more, which benefits worker and employer alike.

Independently of the fixed element, and given the same rate of pay for each piece produced, individual disparities in pay are inevitable. They are even inherent in remuneration at piece-rates, since they reflect differences in output. That is, moreover, clearly illustrated by the fact that, even within each of the groups A and B, there are significant individual differences. ( 43 ) Workers receive equal pay only if their productivity is the same.

Hence, if the fixed element of hourly earnings on a piece-work basis is identical for work of equal value, average earnings do no more than reflect differences in output. That leads me to conclude that, in a pay system of that type, they cannot possibly serve as a relevant criterion for comparison leading to a presumption of the existence of sex discrimination.

65.

I would make one last point, one also mentioned by the German Government. ( 44 ) My views are supported by subparagraph a) of the third paragraph of Article 119 since the only deciding factor in comparisons it mentions is the ‘unit of measurement’ of pay. Discrimination might be found to exist if the ‘unit of measurement’ of the fixed element and/or the ‘unit of measurement’ of the variable element were not identical for work which was of equal value or regarded as such. In other words, discrimination would only arise where, in relation to pay for a unit of time or pay for a piece, a distinction was made between men and women.

66.

By the second part of the third question, the arbitration board asks whether, on the basis of successive comparisons involving groups A and B, their pay may be aligned with that of the group with the highest earnings (group C).

67.

Evaluation of such a situation requires consideration of certain specific facts which have not been put before the Court. What is more, that examination of the facts falls within the exclusive competence of the national court. I shall simply point out that, if such a comparison were to be allowed, the effect would be to align the pay of one group of women (group B) with that of another group of women (group C), whereas the sole purpose of Community provisions on the principle of equal treatment of men and women was, of course, to eliminate all discrimination between men and women.

Consequently, such an approach falls outside the scope of both the Directive and Article 119 of the Treaty.

68.

Above all, however, that question illustrates yet again the risk in disregarding the specific nature of remuneration at piece-rates. If the earnings of the various groups mentioned in the second part of the third questions were indeed raised, the result would be the establishment of a single level of pay for all the groups of workers, irrespective of individual output; it would no longer be individual remuneration at piece-rates.

69.

The fourth question, by which the arbitration board seeks to ascertain the significance attributable to a number of facts, calls for only brief comment.

70.

It is for the arbitration board to consider the facts before it in order to determine the question of equal value in relation to the groups compared. If it decides that, independently of the average earnings, there are significant facts creating a presumption of discrimination, those findings of fact may also help it to determine whether there is objective justification unrelated to any sex discrimination. I would refer in that connection to my earlier remarks regarding the case-law. ( 45 )

71.

I therefore propose that the Court rule as follows: