BAILII [Home] [Databases] [World Law] [Multidatabase Search] [Help] [Feedback]

Court of Justice of the European Communities (including Court of First Instance Decisions)


You are here: BAILII >> Databases >> Court of Justice of the European Communities (including Court of First Instance Decisions) >> Banca popolare di Cremona (Taxation) [2006] EUECJ C-475/03 (03 October 2006)
URL: http://www.bailii.org/eu/cases/EUECJ/2006/C47503.html
Cite as: [2006] EUECJ C-475/3, [2007] 1 CMLR 31, [2006] EUECJ C-475/03, [2006] ECR I-9373

[New search] [Printable version] [Help]


IMPORTANT LEGAL NOTICE - The source of this judgment is the web site of the Court of Justice of the European Communities. The information in this database has been provided free of charge and is subject to a Court of Justice of the European Communities disclaimer and a copyright notice. This electronic version is not authentic and is subject to amendment.



OPINION OF ADVOCATE GENERAL

STIX-HACKL

delivered on 14 March 2006 (1)

Case C-475/03

Banca Popolare di Cremona

v

Agenzia Entrate Ufficio Cremona

(Reference for a preliminary ruling from the Commissione Tributaria di Cremona)

(VAT - Internal taxation - Regional tax on production)





Introduction

The proceedings

1. In the present case, the Commissione Tributaria Provinciale di Cremona (Cremona Regional Tax Court) asks whether Article 33(1) of the Sixth VAT Directive (2) precludes the levying of a tax such as the Italian imposta regionale sulle attività produttive, known more generally by its acronym "IRAP". In the main proceedings, Banca Popolare di Cremona ("Banca Popolare") seeks reimbursement of various sums which it paid by way of IRAP in 1998 and 1999.

2. Following written and oral submissions from Banca Popolare, the Italian Government and the Commission, Advocate General Jacobs delivered his Opinion on 17 March 2005, (3) concluding that a national tax which

- is levied on all natural and legal persons who regularly carry on an activity with the object of producing or trading in goods or providing services,

- is imposed on the difference between the proceeds and costs of the taxable activity,

- is charged in respect of each stage in the production and distribution process corresponding to a supply or set of supplies of goods or services made by a taxable person, and

- imposes a burden at each of those stages which is globally proportional to the price at which the goods or services are supplied

is to be characterised as a turnover tax prohibited by Article 33(1) of the Sixth Directive.

3. However, because the need to reimburse large sums of tax levied contrary to Community law might very seriously disrupt regional funding in Italy, and because the Commission appeared to have contributed by its conduct to the Italian Government's belief that IRAP was compatible with Community law, he also recommended that the Court should fix a temporal limitation on the effects of its ruling.

4. Moreover, in view of various tactics that might be adopted in anticipation of that ruling, he envisaged the possibility of a new approach to such limitation. He pointed out that some national courts may find a measure unlawful while fixing, in order to allow sufficient time for new legislation to be enacted, a future date before which individuals may not rely on that unlawfulness in any claims against the State. However, if such a course of action were to be contemplated in the present case, he thought it desirable for the Court to hear further argument on the point. Seven Member States subsequently requested a reopening of the oral procedure for that purpose.

5. On 21 October 2005, the Grand Chamber decided to reopen the oral procedure, fixed a new hearing for 14 December 2005, and asked the parties to the main proceedings, the Member States, the Council and the Commission to express their views on the following questions: (4)

(a) What are the criteria for classifying a tax as a turnover tax within the meaning of Article 33(1) of the Sixth Directive, having regard to the purpose of that provision and the functioning of the market?

(b) To what extent are bank transactions capable of being subject to such a tax?

(c) Regard being had to Advocate General Jacobs's Opinion, in what circumstances and in what way may the effects of a preliminary ruling delivered by the Court be subjected to a temporal limitation?

6. Banca Popolare, 13 Member States and the Commission submitted written responses, though only a few Member States addressed the first or second questions; and Banca Popolare, 12 Member States and the Commission presented oral submissions at the second hearing, the Member States again focussing for the most part exclusively on the aspect of temporal limitation.

The characteristics of IRAP

7. As Advocate General Jacobs pointed out in his Opinion, only the Italian courts are competent to assess the precise features of IRAP. The role of this Court is to interpret Community law in such a way that the referring court can usefully apply it to the tax which it has to assess. (5) In doing so, this Court must therefore assume the nature of that tax to be as described in the order for reference.

8. That description is worded as follows:

"(1) It flows from the definition in Article 2 [(6)] that IRAP applies, in general, to all commercial transactions involving production or exchange of goods and services arising through the regular exercise of an activity intended for that purpose, that is, through undertakings, trades and professions.

There is therefore an exact correlation between the "precondition of the (IRAP) tax" set out in Article 2 and the area of "taxable transactions" laid down in Article 1 of the decree introducing VAT, which constitutes the precondition of that tax.

(2) Under the first paragraph of Article 4, IRAP taxes the net value deriving from production, that is, the net value "added" to the product by the producer.

IRAP is therefore a tax on the "added value" produced, and indeed the term "vap" [(7)] is used to denote the basis of assessment of the new tax.

Even the taxable subject-matter of IRAP is exactly the same as that of VAT. With VAT, the fraction or portion of added value produced by an individual producer is quantified and assessed via the mechanism of deduction of tax from tax (output tax paid on purchases is deducted from input tax collected on sales). With IRAP, the fraction is calculated and assessed, roughly, by deducting the cost of acquiring the "thing sold" from the proceeds of the "sale".

The fractional assessment of VAT and IRAP are as similar as two drops of water, notwithstanding the difference in the technical devices used for calculating the taxable amount and the tax itself. To establish how much VAT an individual trader must pay, the VAT on the cost of sales is deducted from the VAT on sales; if the difference produces a positive figure, it is the VAT due to the tax authorities, and from this the taxable value added by the trader can be quantified.

The process is reversed for IRAP.

Instead the added value produced is the starting point for obtaining the amount of tax due rather than the other way around. This difference does not affect the essence of the question which is this: both taxes tax, at all stages of the production and distribution process, the fraction of value added by the individual producer involved in the production and/or distribution process. Whereas VAT deducts tax from tax, IRAP deducts basis from basis, costs from amounts due.

(3) IRAP is levied at every stage of the production or distribution process, since every trader involved in a stage of the cycle, producing taxable added value, is considered by the law to be liable to tax. Suppose that there are three stages of the cycle, depending on Tom, Dick and Harry; all three are separately and independently liable to IRAP, each assessed out of 100. The same occurs with VAT.

(4) Finally it should be noted that the amount of IRAP collected in the various stages of the cycle, from production up to the final consumer, is equal to the rate of IRAP applied to the selling price of goods and services charged to the final consumer. Despite its fractional basis, therefore, IRAP in fact acts as a general and proportional tax on the price of transferring goods or services to the consumer."

9. Advocate General Jacobs based his analysis on that description, and on findings made by the Italian Corte Costituzionale (Constitutional Court) in judgment 256/2001 of 10 May 2001, in which it dismissed various challenges to IRAP on grounds of alleged incompatibility with the Italian Constitution.

10. In particular, he noted, the Corte Costituzionale had held that IRAP "is not levied on the taxpayer's personal income but rather on the added value produced by independently run activities" and that "the economic burden of the tax may in fact be passed on in the price of the goods or services produced, according to the laws of the marketplace, or be wholly or partly recouped by means of appropriate organisational choices". (8)

11. To the extent that any aspect of the national courts" description of IRAP may differ from that in the submissions to the Court - and Banca Popolare, the Italian Government and the Commission have all provided their own descriptions - the Court must in principle accept the account given in the order for reference. However, the existence and nature of any such differences may indicate areas in which it may be helpful for the Court to consider also how the answer it gives might apply to possibly slightly different features. In any event, if the referring court's assessment of the characteristics of IRAP is disputed in the main proceedings, its judgment will presumably be subject to appeal on that ground within the national judicial order.

The assessment

12. My assessment of the case will accordingly be in two parts: a first part dealing with the compatibility of a tax such as IRAP, and a second part dealing with the possibility of a temporal limitation of the effects of the judgment.

13. Within each of those parts, I shall set out first a number of general considerations, including a review of the Court's case-law, then I shall seek to apply those considerations to the circumstances of the present request for a preliminary ruling.

I - The compatibility of a tax such as IRAP with the Sixth Directive

A - General considerations

1. Taxes prohibited by Article 33(1) of the Sixth Directive

14. There is a well-established body of case-law concerning the circumstances in which a national tax falls foul of the prohibition of levies other than those "which cannot be characterised as turnover taxes" in Article 33(1) of the Sixth Directive. (9) In particular, there are a number of specific criteria in that case-law which the national court evidently had in mind in drafting its order for reference, and against which Advocate General Jacobs assessed IRAP as described in that order. (10)

15. The fact that on reopening the oral procedure the Court has asked the parties, the Member States and the institutions for their views on the criteria for classifying a tax as a turnover tax within the meaning of Article 33(1), having regard to the purpose of that provision and the functioning of the market, might suggest that it envisages the possibility of reviewing, refining or developing those criteria. I shall therefore consider them in some detail.

2. Synthesis of the existing case-law

16. The Court's previous case-law in this field comprises in particular (11) the judgments in Rousseau Wilmot, (12)Bergandi, (13)Wisselink, (14)Giant, (15)Dansk Denkavit, (16)Bozzi, (17)Beaulande, (18)Careda, (19)UCAL, (20) Solisnor, (21) SPAR, (22)Pelzl, (23)EKW, (24)Tulliasiamies (25) and GIL Insurance. (26) Of those cases, it may be noted that only Bergandi, Wisselink and Dansk Denkavit were decided by a plenary formation of the Court, the remainder being decided by chambers.

17. In that case-law spanning a period of nearly 20 years, a number of strands, and a certain development, can be discerned.

18. First, the Court has had regard to the objectives of the VAT legislation and to the purpose of the prohibition of other types of turnover tax. Next, it has defined a number of characteristics of VAT, possession of which brings another national tax within the scope of the prohibition. Finally, in the course of examining the various national taxes in issue, it has identified various specific characteristics which are or are not permissible in that regard, together with a number which it has not regarded as relevant.

19. Three principal objectives are identified for the adoption of the common system of VAT:

- to abolish and replace cumulative multistage turnover taxes as formerly applied in many Member States; (27)

- to establish a common market within which there is healthy competition and whose characteristics are similar to those of a domestic market, by eliminating tax differences liable to distort competition and hinder trade; (28) and

- to ensure equal conditions of taxation for a given transaction, regardless of the Member State in which it takes place. (29)

20. The purpose of the prohibition of other turnover taxes is stated in many judgments (30) to be to prevent the functioning of the common system of VAT from being jeopardised, and it is specified that such a result would ensue from the imposition of taxes, duties or charges levied on the movement of goods and services in a way comparable to VAT. In Wisselink, (31) a further more general reference is made to jeopardising the objectives underlying the common system of VAT at the current stage of harmonisation.

21. The characteristics of VAT are described initially by reference to Article 2 of the First Directive and to Article 17(2) of the Sixth Directive: the principle consists in the application to goods and services up to the retail stage of a general tax on consumption which is exactly proportional to the price of the goods and services, irrespective of the number of transactions which take place in the production and distribution process before the stage at which the tax is charged; however, VAT is chargeable on each transaction only after deduction of the amount of VAT borne directly by the costs of the various price components, taxable persons being authorised to deduct from the tax for which they are liable the tax which the goods have already borne. (32)

22. Then, from Dansk Denkavit onwards, there is a more formal list of "essential characteristics" of VAT, namely:

- it applies generally to transactions relating to goods or services;

- it is proportional to the price charged by the taxable person in return for the goods and services which he has supplied;

- it is charged at each stage of the production and distribution process, including that of retail sale, irrespective of the number of transactions which have previously taken place;

- the amounts paid during the preceding stages of the process are deducted from the tax payable by a taxable person, with the result that the tax applies, at any given stage, only to the value added at that stage and the final burden of the tax rests ultimately on the consumer. (33)

23. The characteristics of a national levy which is caught by the prohibition in Article 33(1) of the Sixth Directive are described, first, in general terms as those which have the effect of compromising the functioning of the common system of VAT by burdening the movement of goods and services and commercial transactions in a way comparable to VAT. (34) In those judgments in which the four essential characteristics of VAT are listed, (35) it is specified that taxes, duties and charges must in any event be regarded as being imposed on the movement of goods and services in a way comparable to VAT if they exhibit those essential characteristics. It is not however necessary for them to resemble VAT in every respect. (36)

24. Then, at a more specific level, the Court has only once found a tax to be actually incompatible, in Dansk Denkavit, and it described the features of that tax as follows: (37)

- it was paid both on activities subject to VAT and on other industrial or commercial activities consisting in the supply of services for consideration;

- it was charged, in the case of undertakings which were taxable persons for VAT purposes, on the same basis of assessment as that used for VAT, in other words as a percentage of the volume of sales after deduction of purchases;

- unlike VAT, it was not paid on importation, but was charged on the full sale price of imported goods at the first sale in the Member State concerned;

- unlike VAT, it did not have to be indicated separately on invoices; and

- it was charged alongside VAT.

25. In Careda, (38) the Court further specified that in order to be caught by the prohibition, a tax must be capable of being passed on to the consumer.

26. In the remaining judgments, however, the Court has found various national levies to be compatible with the Sixth Directive because their characteristics differed sufficiently from those of VAT. In most cases, the Court has based its finding on possession of a bundle of characteristics distinguishing the tax from VAT, but in Solisnor (39) and EKW (40) it considered that a lack of general application was sufficient in itself to exclude the tax from the prohibition; a limitation to certain categories of goods or services has also been taken into account in most of the other judgments and seems in any event to have been a feature of all the disputed levies, except perhaps those in Rousseau Wilmot and SPAR.

27. Other characteristics which the Court has found to differ from those of VAT, and to be relevant when assessing the compatibility of a national tax, are:

- calculation of the tax on the basis of annual turnover, so that it is impossible to determine the precise amount passed on to customers; (41)

- imposition on the basis that an article is made available to the public, regardless of the amount or even the existence of a charge for use, or calculation on the basis of a deemed rather than an actual price; (42) more generally, calculation on a basis other than that of added value; (43)

- absence of any provision for deduction of input tax; (44)

- lack of direct or strict proportionality to the price of the transaction burdened; (45)

- imposition at only one stage in a chain of transactions; (46) however, in Wisselink, (47) the Court made it clear that a tax levied once only could still be contrary to Community law if it prevented the proper working of the VAT system;

- imposition on the taxable person's input, and not output, transactions. (48)

28. Finally, the Court has also found a number of characteristics not to be relevant when assessing the compatibility of a national tax:

- the designation of the tax in national law, the express wording of the legislation or the reasons for its adoption; (49)

- the fact that the tax is levied concurrently with VAT; (50)

- the absence of a specific requirement that the tax be passed on to customers; (51)

- the absence of a requirement that the tax be mentioned on an invoice or that an invoice be issued or held; (52)

- the fact that the tax is not paid on importation but on the first supply thereafter; (53)

- the fact that the tax uses a different basis of assessment when applied to taxable persons not subject to VAT. (54)

3. General comments on the case-law

29. There are thus two broad aspects to the Court's approach: a concern, at a fundamental level, to protect the principles underpinning the VAT system and to preclude interference with it; and a desire to define, more formally and in the interests of legal certainty, criteria whereby national taxes may be clearly identified as incompatible with the VAT system.

30. A survey of the case-law indicates that, while the first concern has never been overlooked, the reference to specific criteria has acquired increasing importance in more recent judgments. Those Member States which have responded to the Court's first question (55) have all favoured confirming that importance; Finland in particular pleads for firm adherence to the four "essential characteristics" in the interests of clarity, consistency and legal certainty.

31. That position is understandable. Member States must know what limits may apply to their freedom of action in introducing new forms of taxation or in maintaining or varying existing forms. A set of clear, objective criteria is important in that context.

32. Yet there is always a danger that a purely formal application of any rules or criteria could lead to results at odds with the fundamental purpose envisaged when those rules or criteria were formulated - in the present case, that of ensuring that the proper functioning of the common system of VAT is not jeopardised.

33. It therefore seems to me indispensable, when assessing the features of any disputed national levy against the essential characteristics of VAT, to continue to do so in the light of that purpose, with a view to preventing any interference with the basic objectives of replacing cumulative multistage turnover taxes, providing healthy competition in the internal market, eliminating tax differences liable to distort competition or hinder trade, and ensuring equal conditions of taxation for a given transaction, regardless of the Member State in which it takes place.

34. In that context, the characteristics of a national tax may vary in their degree of similarity with the essential characteristics of VAT. It seems unlikely that, if a tax jeopardises the functioning of the common system because it possesses those characteristics in an identical form, it will cease to jeopardise the system simply as a result of minor differences.

35. The Court has rightly taken the view that a tax need not be identical to VAT in all respects in order to be caught by the prohibition, but must in any event be regarded as being caught if it exhibits the essential characteristics of VAT. By the same token, I consider, what is necessary for each of those individual characteristics is not strict and absolute identity but essential, substantive identity.

36. Conversely, it might conceivably be the case that in practice, and perhaps paradoxically at first sight, a tax could interfere less with the VAT system as it resembles it more. A hypothetical tax additional to VAT but otherwise identical to it in all respects would differ little in effect from an increase in the standard rate of VAT, for which there is at present no upper limit. (56) What is likely to interfere most with the common system is a tax which, whilst possessing essential features of VAT, also possesses features which conflict with it. (57)

37. In that connection, Advocate General Jacobs accepted (58) that possession of all four essential features of VAT is a necessary condition for a national tax to be incompatible with the Sixth Directive, which would imply that in the absence of any one of those features a tax will be compatible.

38. A slightly different interpretation of the case-law might also be possible.

39. In all cases but Solisnor and EKW the Court has stressed that the national tax in question lacked more than one of the four essential characteristics; and in both those cases it was the absence of general application which was decisive. That might suggest that, of the four characteristics, that of general application is to be considered as having greater weight than the other three. However, it does not seem to me that such a conclusion can be drawn with any certainty, and I do not propose to adopt that interpretation.

40. What then can be said, in general terms and with confidence, about the essential characteristics of VAT, viewed in the light of the purpose of the common system and of the prohibition of national taxes likely to jeopardise its functioning?

41. At least a number of negative points can be made. A tax is not likely to jeopardise that functioning if it does not apply generally; taxes confined to specific categories of goods or services are not likely to interfere with the system as a whole. A tax which is not levied at each stage in the production or distribution chain is less likely to jeopardise the system; such taxes may affect a particular stage in the chain, but not the whole system. (59) A tax which is not proportional to the value added at each stage, and thus to the overall price at each stage, is less likely to interfere with the VAT system; flat-rate taxes can in general coexist in parallel with proportional taxes. (60) And of course a tax which is not capable of being passed on to the consumer cannot compromise the functioning of VAT as a tax on consumption.

42. Before examining the application of the case-law, and of the criteria which it enunciates, to a tax such as IRAP, however, and still at a general level, it is necessary to consider the relevance, if any, of two specific points: the status of bank transactions in relation to turnover taxes, and the distinction between direct and indirect taxation in the light of Article 93 EC.

4. Subjection of bank transactions to turnover tax

43. The Court asked for the views of the parties, the Member States and the Commission on the extent to which banking transactions are capable of being subject to a turnover tax within the meaning of Article 33(1) of the Sixth Directive.

44. Banca Popolare, the French and Hungarian Governments and the Commission replied briefly, all agreeing that banking transactions may in principle be subject to such a tax (although there might be greater practical difficulties than in the case of other commercial transactions) but also stressing the broad exemptions from VAT for financial services. (61)

45. The Hungarian Government considers as a result that a Member State is not precluded by Article 33(1) from imposing a tax other than VAT on banking transactions. Banca Popolare disagrees, stressing that the referring court's question concerns the compatibility of IRAP "tout court", and not of its application to banking transactions.

46. I do not think that a deep analysis of this question would prove profitable. (62) Banking services come within the scope of VAT by virtue of Article 6(1) of the Sixth Directive; they are largely exempt under Article 13(B)(d); but Member States may allow taxpayers a right of option for taxation under Article 13(C)(b). If a tax of general application meets all the criteria to be caught by the prohibition in Article 33(1), it should not fall outside that prohibition simply because it applies also to a category of transaction which is exempt and for which, as the case may be, the Member State in question has not provided for an option for taxation. (63)

47. It does not therefore seem to me relevant to the Court's assessment that most of the transactions carried out by Banca Popolare, or by banks in general, are likely to be exempt from VAT.

5. Article 93 EC and the distinction between direct and indirect taxes

48. Finally, I turn to a general point raised by a number of Member States in their observations, namely whether the possible classification of a tax as a direct tax may be relevant to assessment of its compatibility with the Sixth Directive.

49. All the VAT directives are based on Article 93 EC (formerly Article 99 of the EC Treaty), under which the Council is to adopt provisions for the harmonisation of legislation concerning "turnover taxes, excise duties and other forms of indirect taxation" to the extent "necessary to ensure the establishment and the functioning of the internal market" (originally "in the interest of the common market").

50. It has been argued therefore that since Article 93 EC concerns only indirect taxation, the Sixth Directive cannot prohibit any levy which is not itself an indirect tax.

51. In my view, however, that conclusion is not warranted.

52. First of all, it may be noted that, as the Commission pointed out at the second hearing, the Court has never categorised a national tax as "direct" or "indirect" when assessing its compatibility with the Sixth Directive. (64) Clearly, no firm conclusion can be drawn from that fact, but it does indicate that the distinction between direct and indirect taxation has not as such been regarded hitherto as an essential criterion, and might suggest that it should be introduced as a criterion only if there are strong grounds for doing so, and it remains to be seen whether that is the case.

53. In that connection, it should be recalled that the Treaty contains no definition of a difference between direct and indirect taxation, and it is commonly accepted that no complete, unambiguous and universally valid definition can be formulated.

54. Certainly, the essence of the distinction is clear: a direct tax is collected directly from the person who bears the economic burden; an indirect tax is included in a sum paid by that person to another, who does not bear the economic burden but who accounts for the tax.

55. Typical direct taxes include those on personal wealth, property or income; and it may be the case that only taxes levied on individuals as such are wholly direct in nature. By contrast, VAT as regulated by the Sixth Directive is an indirect tax par excellence, since it is completely neutral as regards the economic operators from whom it is collected and is in principle always passed on to the final consumer as an identifiable proportion of the price (and as a specified sum when an invoice is issued).

56. However, the situation is not always as clear-cut. Some taxes which are primarily direct may partake of the nature of indirect taxation, and vice versa. And even if a satisfactory criterion could be found for distinguishing in law between direct and indirect taxation, some of the effects of a tax such as IRAP seem capable of overlapping with those of VAT, so that interference between the two cannot be ruled out.

57. Next, it has been rightly pointed out that Article 93 EC clearly cannot provide a valid legal basis for Community harmonisation of direct taxation. Yet, just as clearly in my view, it can provide such a basis for Community legislation prohibiting national taxation likely to jeopardise the functioning of a form of harmonised indirect taxation - such as VAT. It does not seem to me necessary to require a different legal basis simply because the national levy in issue may have at least some characteristics of direct taxation. What matters is whether it possesses features which are likely to jeopardise the functioning of the VAT system, regardless of whether it also possesses others which are not - and it should be remembered that only offending features will fall foul of the prohibition based on Article 93 EC.

58. Put more succinctly, Article 93 EC does not allow Community legislation to encroach on Member States" fiscal sovereignty in the field of direct taxation; conversely, it does not allow Member States to adopt, in their exercise of that sovereignty, measures likely to jeopardise agreed harmonisation of indirect taxation. Consequently, it may serve as a valid legal basis for legislation which precludes such measures.

59. Finally, it may be noted that although since 1989 (65) all VAT directives appear to have been adopted on the basis of Article 93 EC (or its predecessor Article 99 of the EC Treaty) alone, that was not true for earlier directives. Until 1986 (66) they were based on both Article 99 and Article 100 of the EC Treaty, that is to say the current Articles 93 and 94 EC. Under the latter, the Council is to issue directives for the approximation of such national provisions "as directly affect the establishment or functioning of the common market".

60. Thus the Sixth Directive, being among the earlier group, was adopted on the basis not only of a Treaty provision which authorises harmonisation of indirect taxes but also of another provision which authorises more generally all harmonisation in direct relationship with the establishment and functioning of the common market.

61. Consequently, there is in my view no reason why the prohibition of national turnover taxes other than VAT may not extend to all taxes which are capable of jeopardising the functioning of the VAT system. That depends on the features and effects of each tax, and not on a theoretical assignment to the class of direct or indirect taxation. There do not appear to be any cogent grounds for introducing such a criterion into the assessment of compatibility with Article 33(1) of the Sixth Directive, and Article 93 EC, as the legal basis of the Sixth Directive, is therefore of no direct assistance in this regard.

6. Conclusion on the general considerations concerning compatibility with the Sixth Directive

62. I thus reach the view that neither the specificity of bank transactions nor the question of classification as direct or indirect taxation need influence the assessment in this case, which should be based on an examination of the four essential characteristics all of which must be present for a national tax to fall within the prohibition in the Sixth Directive.

63. However, it seems to me that the Court faces, with regard to certain aspects of those criteria, a choice between a narrow application and a broader application, the latter being at least implicitly inherent in Advocate General Jacobs's analysis.

64. And it may be seen from my outline of the case-law that both the broader and the narrower approach are already present, to varying degrees, in the Court's existing case-law, which may thus have maintained a degree of uncertainty. In the present case, the Court must dispel any such uncertainty by a clear indication as to whether the four essential characteristics, used as criteria for determining whether a national tax is prohibited by the Sixth Directive, are to be viewed purely formally or in the light of the objectives of the prohibition and of the harmonised VAT system as a whole.

65. Given moreover the status of this case - the attention which it has attracted, the fact that, unlike most of the previous cases, it is to be decided by a Grand Chamber of the Court and the fact that the procedure has been reopened specifically to consider, inter alia, the criteria for classifying a tax as a turnover tax within the meaning of Article 33(1) of the Sixth Directive - it must also be borne in mind that the choice made will determine decisively the course to be followed in future judgments.

B - Application to a tax such as IRAP

66. I shall now examine how the criteria developed in the Court's case-law should be applied to IRAP, bearing in mind that it is not for this Court to define the characteristics of that tax. That is for the Italian courts, and the description in the order for reference should be taken as authoritative as a matter of principle. However, consideration may be given to aspects of that description which appear to be contested.

67. Advocate General Jacobs did, of course, examine the same criteria in his Opinion, and there would be little to be gained by repeating his analysis.

1. Generality of application

68. The first essential characteristic of VAT according to the Court's case-law is that it applies generally to transactions relating to goods and services.

69. According to the order for reference, "IRAP applies, in general, to all commercial transactions involving production or exchange of goods and services arising through the regular exercise of an activity intended for that purpose" and the "fractional assessment of VAT and IRAP are as similar as two drops of water, notwithstanding the difference in the technical devices used for calculating the taxable amount ...".

70. Advocate General Jacobs dealt with this aspect at points 28 to 40 of his Opinion, and concluded that IRAP possessed the characteristic in question.

71. In that light and in the light of paragraph 15 of the judgment in Dansk Denkavit, the general application of IRAP to goods and services does not seem to be an issue: it is undisputed that, unlike almost all of the national taxes examined by the Court and found to be compatible with the Sixth Directive, the application of IRAP is not limited to particular categories of goods or services.

72. Further, a difference of opinion between Banca Popolare and the Italian Government, to which Advocate General Jacobs referred at point 32, seems to me to have been clarified at the second hearing. Banca Popolare denied that IRAP was levied on goods produced but not yet sold, thus burdening stocks and not only supplies. If I understood correctly at the hearing, it seems that the Government's statement is strictly true but that since stocks are valued at cost price and the cost of production is deducted from the tax base the effect is that in almost all such cases the amount of tax levied is actually zero. Consequently, that issue need not detain the Court either.

73. There remains however an aspect on which it may be necessary to consider whether a broader or a narrower view should be taken: despite the wording of the order for reference, it appears that IRAP is not calculated on individual transactions as such, whereas the Court's case-law, inspired by Article 2 of the First Directive, defines VAT as applying generally to "transactions". (And, as will become clear, that distinction is relevant to more than one of the essential characteristics of VAT.)

74. Advocate General Jacobs concluded that in practice there is little difference between a tax calculated, like IRAP, on the basis only of a periodic aggregate and one calculated, like VAT, on a transaction-by-transaction basis for individual invoices and on the basis of a periodic aggregate for traders.

75. It does indeed seem evident that the difference between total inputs and total outputs over a tax period must be the same as the sum of the differences between individual inputs and individual outputs over that same period.

76. The Court could none the less apply its previous case-law strictly, considering that only levies calculated on individual transactions as such are caught by Article 33(1) of the Sixth Directive, regardless of periodic aggregation. In a number of judgments (67) the fact that a tax is calculated on the basis of annual turnover has been considered a relevant characteristic distinguishing that tax from VAT, although there does not seem to be any instance in which no other distinguishing characteristics have been present.

77. Such an approach would make incompatibility conditional on a more exact correspondence than that suggested by Advocate General Jacobs between the manner of calculating the national tax and that of calculating VAT.

78. It would also imply a change in the case-law as compared to Dansk Denkavit, delivered by a plenary formation of the Court: the tax found to be incompatible in that judgment was calculated according to a mechanism apparently very similar to that of IRAP, being charged "as a percentage of the volume of sales after deduction of purchases".

79. In my opinion, the broader view taken in Dansk Denkavit is the more appropriate, encompassing within the prohibition taxes which, although not calculated in precisely the same manner, lead to the same economic result for traders and consumers as those imposed on individual transactions. Such a view better serves the aim of ensuring that the functioning of the VAT system is not affected.

80. Taking that view, and on the assumption that the method of calculation of the tax in question does give rise in practice to the same amount of tax due as a method based on the values of individual transactions (regardless of periodic aggregation), it seems to me clear that the characteristic of general application to transactions relating to goods and services is present in a tax such as IRAP.

81. It is true that the Italian Government denies that IRAP, calculated on the overall results of a business, can be likened to VAT calculated on each of the individual transactions carried out by that business.

82. However, its arguments are not based directly on the difference between the methods of calculation. Rather, it submits, first, that IRAP is imposed on the business as such and not on its transactions, so that it is a direct and not an indirect tax; (68) that is an aspect I have dealt with above at point 48 et seq., where I concluded that it was not a relevant consideration. Second, in the Italian Government's submission, the aggregate value on which IRAP is based is not equivalent to the sum of the individual values on which VAT is based; that however is a submission concerning the nature of the added value used as tax base rather than the method of calculating the amount due, and I shall consider it in the context of the fourth essential characteristic, (69) where it is more relevant than here.

83. Those considerations do not therefore affect my view on the assessment of a tax such as that described by the referring court, in relation to the first essential characteristic of VAT.

84. Before turning to the second of those characteristics, I would mention briefly a submission made by the Hungarian Government, to the effect that a tax is not of general application if it is levied at a local or regional level, in particular if it is optional and/or the local or regional authority may determine the rate levied. That does not seem to me to be a relevant criterion; it was certainly not considered relevant by the Court in Pelzl or EKW, both of which involved regional taxes. (70) The question is rather whether the tax is general within its area of application, be that area national or merely local.

85. I am consequently of the view that a national tax having the features described by the referring court shares with VAT the essential characteristic of general application to transactions relating to goods and services. If however, the Court were to decide - contrary to its decision in Dansk Denkavit - that the essential characteristic in question necessarily involves application to individual transactions, such a tax would not possess it.

2. Proportionality to price

86. The second essential characteristic of VAT is that it is proportional to the price charged by the taxable person for the goods or services he supplies.

87. The order for reference states that "the amount of IRAP collected in the various stages of the cycle, from production up to the final consumer, is equal to the rate of IRAP applied to the selling price of goods and services charged to the final consumer. ... [T]herefore, IRAP in fact acts as a general and proportional tax on the price of transferring goods or services to the consumer".

88. Advocate General Jacobs dealt with this aspect at point 59 et seq. of his Opinion. He accepted that the mechanism of IRAP was such that it might not always remain strictly proportional to the price of each transaction, but considered that aspect not to be conclusive since, in essence, the effective burden of VAT, as opposed to the amount shown on the invoice, may be redistributed by a trader amongst the various types of supply he makes, by means of an adjustment of profit margins, in the same way as may be done for IRAP. Or, put another way, both taxes are proportional in nature and traders will, in the normal course of events, distribute the burden of each proportionally among their products and services; in either case they may depart from that distribution but will only do so in order to achieve a particular commercial purpose.

89. It is true that in a number of cases (71) the Court has considered that a lack of strict proportionality, or a method of calculation which did not allow the precise amount of tax passed on to the consumer to be determined, was a relevant distinguishing characteristic, at least when taken with others, of a levy not prohibited by Article 33(1) of the Sixth Directive. If that case-law is followed, it seems to me that IRAP may also be distinguished from VAT on that ground.

90. However, the tax found to be incompatible in Dansk Denkavit was described by the referring court in that case as "charged at each stage as a percentage of the undertaking's sales ... with deduction of purchases on which the levy has been paid at an earlier stage", and not indicated separately on invoices - which seems also to be the situation with IRAP. And in Careda (72) the Court clearly decided that it was sufficient for a tax to be capable of being passed on to the consumer but unnecessary for there to be any requirement to that effect; in such circumstances it is obviously not possible to apply also the condition that the amount of tax must always be ascertainable as a constant proportion of the price of each transaction.

91. The Court appears therefore to have varied between a narrower and a broader approach to the criterion of proportionality to price and in the present case it must decide which of those approaches is more appropriate.

92. It seems to me that the broader approach is required if the aim of not jeopardising the functioning of the VAT system is to be kept in sight. It does not seem possible to consider that a tax interferes with that system if it is strictly proportional to transaction prices but not if it is only broadly proportional. As Advocate General Jacobs pointed out in essence at point 69 of his Opinion, that would enable a Member State to escape the prohibition by a perhaps very minor adjustment whilst none the less levying a tax which substantially overlaps with VAT in its operation.

93. Another point which may however be considered in this context and which has been stressed by the Italian Government is the treatment of imports under IRAP. Although not specified in the order for reference, it appears to be common ground that IRAP burdens only the value added within (the relevant region of) Italy. Consequently, if an item is imported for (processing and) resale, the total amount of IRAP levied will be proportional only to the difference between the import cost and the selling price, and not to the whole of the selling price as would be the case for VAT.

94. That feature does indeed distinguish IRAP from VAT and from the tax in Dansk Denkavit, the former being applicable to the value of imports as such and the latter to that value as incorporated in a subsequent supply. On a strict view, therefore, it is a feature capable of excluding IRAP from the prohibition.

95. However, again I am not convinced that it does so on a broader view which takes account of the aim of preventing the functioning of the VAT system from being jeopardised.

96. It is in the nature of national taxes - at least in the case of indirect taxes - that if imposed on added value they will not be imposed on value added outwith their fiscal jurisdiction; and a tax on imports as such would fall foul of Articles 25 and 26 EC. A criterion of proportionality to price in the context of a prohibition of national taxes which interfere with the operation of the common VAT system might therefore be taken to mean proportionality with regard to the value added to the price within the fiscal jurisdiction of the Member State concerned.

97. On this criterion, therefore, I am of the view that IRAP as described is sufficiently proportional to the price charged on the transactions which it burdens not to be excluded from the prohibition in Article 33(1) of the Sixth Directive if that prohibition is viewed as precluding national taxes which overlap and thus interfere with the operation of VAT. That would not be the case if the criterion were to be applied in a more formal manner.

3. Charging at each stage

98. The third essential characteristic of VAT is that it is charged at each stage of the production and distribution process, up to and including that of retail trade, whatever the number of transactions which have already taken place.

99. The order for reference states that "IRAP is levied at every stage of the production or distribution process, since every trader involved in a stage of the cycle, producing taxable added value, is considered by the law to be liable to tax".

100. Advocate General Jacobs dealt with this aspect at points 55 to 58 of his Opinion, concluding that the criterion was met by that description.

101. The principal objections to such a view are similar to those I have considered above, namely the fact that IRAP is not levied at the stage of actual transactions but on the basis of annual figures, and that it does not burden imports.

102. With regard to the first objection, it seems to me that a strict approach would not be warranted in any event. IRAP is levied, as Advocate General Jacobs put it, "in respect of each stage in the production and distribution process corresponding to a supply or set of supplies of goods or services made by a taxable person". This third "essential characteristic" of VAT can only be taken to mean that there are no stages in the chain of supply which escape it, (73) and it is perhaps significant that most, if not all, of the levies found by the Court not to be incompatible with Article 33(1) of the Sixth Directive have been single-stage taxes. (74) The aim is essentially to protect the functioning of VAT from interference from cumulative multistage taxes, and the criterion of charging at each stage in the chain must be viewed in that light.

103. With regard to the second objection, I would again say (75) that it is inherent in any national tax levied on added value that it is levied in respect of value added within the relevant fiscal territory. Since a tax applicable within the borders of a Member State is clearly capable of jeopardising a common, Community-wide system, it is not decisive whether that tax also applies to events occurring at those borders.

104. Consequently, I agree with Advocate General Jacobs on this point, and see no scope for a narrower application which would exclude IRAP from the prohibition on this criterion.

4. Imposition on added value, with deduction of input tax

105. The last of VAT's four essential characteristics is that it is imposed on the value added to a supply at each stage, with a mechanism for deduction of the tax paid on the value added at previous stages.

106. Advocate General Jacobs dealt with this characteristic at points 41 to 54 of his Opinion, concluding that it was essentially the same for VAT and for IRAP. Again, he took the broad approach which the Court must decide whether to follow or not.

107. I shall deal first with the second, and in my view more straightforward, aspect of this characteristic, namely the existence of a deduction mechanism for input tax.

108. Clearly, IRAP does not possess precisely such a mechanism because it is not calculated, as in the case of VAT, by deducting input tax from output tax, but by deducting input value from output value. What counts is that there is substantive similarity between those two types of mechanism or, as Advocate General Jacobs put it, that the difference between them is limited in practice.

109. In any event, it seems to me, the existence of a deduction mechanism is not in any way an independent criterion, but rather a necessary corollary of the fact that the tax is levied on the value added at each stage and not on the cumulative value.

110. Indeed, possession of a mechanism for deduction of input tax is clearly not itself a characteristic necessary for a tax to be prohibited, since one of the primary express aims of VAT is to replace individual cumulative multistage systems (which by definition lack a deduction mechanism) by a common non-cumulative multistage system.

111. Potentially more problematic is the question of the nature of the added value itself.

112. The referring court states that: "IRAP taxes the net value deriving from production, that is, the net value "added" to the product by the producer. [It] is therefore a tax on the "added value" produced ... the fraction or portion of added value ... is calculated and assessed, roughly, by deducting the cost of acquiring the "thing sold" from the proceeds of the "sale"."

113. On that basis, IRAP would appear to possess the fourth essential characteristic of VAT.

114. The Italian Government has however objected that, while IRAP is indeed imposed on "added value" in an economic sense, it is not the same added value as for VAT. It states that the added value on which IRAP is levied is calculated on the basis not only of costs of input supplies and receipts from output supplies but also of criteria such as variations in stocks (whether or not as a result of supply), variations in value of current work performed, social security contributions, amortisation etc.; and that IRAP allows deduction of imports but burdens exports, whereas VAT is levied on imports but reimbursed on exports. As a result, two firms which account for similar amounts of VAT may pay very different amounts of IRAP.

115. If that is so, it could be deduced that IRAP is not caught by the prohibition on a strict application of this criterion.

116. On the other hand, it might be asked whether the degree of difference is not important. It seems to me that a tax should not escape the prohibition in Article 33(1) of the Sixth Directive simply because the added value on which it is levied is defined in different terms from those of the directive. For it to be caught, there must none the less be a substantive similarity between the two types of added value likely to cause interference and thereby jeopardise the functioning of the common system. There are two levels to the similarity in question.

117. The first is the central principle characterising the tax base, which must be clearly definable as the value added at each stage for there to be substantive similarity. In the case of IRAP, that does not appear to be disputed, and can be acknowledged by the Court in the light of the legislation submitted to it.

118. The second level is the detailed content of the basis of assessment, the elements included in or excluded from the added value taxed. Here, it is common ground that not all the elements relevant for VAT are relevant for IRAP, and vice versa. However, it does not seem to me that a mere enumeration of those elements is of assistance; what is in issue is substantive similarity, and that can be determined only by looking at substantive effects.

119. This Court does not have sufficient information on such effects, nor is it the proper forum for establishing them; a final determination can only be made in the Italian courts. The Court can none the less provide guidance as to criteria whereby a substantive similarity can be objectively identified. I propose a test based on parallelism between the amounts of VAT and of IRAP levied on the value added by each business.

120. If two taxes are calculated at different rates on the same basis of assessment, the ratio between the two will remain constant for every business concerned. One could plot a graph showing the amount of each tax which a number of businesses in different sectors, arranged progressively from the smallest to the largest, had to pay to the tax authorities, and the two lines would be parallel.

121. If one element in the tax base were modified for one of those taxes, that might well affect the parallelism, the constancy of the ratio. The greater the number of elements removed from one tax base and/or added to the other, the more likely it would seem that the parallelism would be affected.

122. Yet that need not always be so. If the basis of assessment for both taxes were added value, but certain elements of that value were present in the case of one tax and absent in the case of the other, the added value in the latter case might still remain a constant though lower proportion of that in the former, so that the parallelism between the two graphs plotted was unaffected.

123. In such a situation, it would in my view have to be concluded that there was substantive similarity between the two notions of added value, since the same effect could be achieved by adjusting the rate of taxation rather than the definition of the tax base. Moreover, it would be clear that the nature of the activity was irrelevant and that strategic choices made by each business - whether to invest in more personnel or more machinery, whether to borrow or to plough back profits, and so on - did not affect that result.

124. If however the differences between the component elements of the two tax bases led to a flagrant lack of parallelism between the two graphs, with random variations, it would in my view have to be concluded that there was no substantive similarity between the taxes.

The same would be true if the variations were not random but could be systematically attributed to strategic choices of the kind exemplified above, having different effects on VAT and the disputed tax.
But in the case of differences between sectors of activity, I would reach that conclusion only if there were also variations within sectors; a parallel graph within each sector would be strongly indicative of substantive similarity of the tax base.

125. Such a test seems to me to be adequately objective even if, of necessity, some of the details must be left to the assessment of the national courts with, if appropriate, the benefit of expert statistical advice. In particular, it will be necessary to determine what constitutes a representative sample of businesses and the extent to which minor variations in the ratio between the amounts of the two taxes can be disregarded. Neither of those should however prove problematic in my view if the principle of the comparison is clear.

126. Subject therefore to a final determination by the Italian courts of the substantive similarity between the added value on which IRAP is based and that on which VAT is based, I am of the view that a national tax having the features described by the referring court shares with VAT the essential characteristic of being imposed on the value added to supplies at each stage, with a deduction mechanism in respect of previous stages.

5. Conclusion on the assessment of a tax such as IRAP in the light of Article 33(1) of the Sixth Directive

127. In the light of the above considerations, I am in broad agreement with Advocate General Jacobs's conclusion as to the incompatibility with Article 33(1) of the Sixth Directive of a tax meeting the description of IRAP given in the order for reference. Of the two approaches I have endeavoured to present, his favoured the broader, and so does mine. It appears to me to be of particular significance in that regard that the broader approach is consistent with the judgment in Dansk Denkavit, decided by a plenary formation of the Court, whereas other judgments which may on some aspects appear to favour a narrower approach have been given by chambers.

128. Consequently, I am of the opinion that a tax having the features of IRAP as described in the order for reference possesses the four essential characteristics of VAT and therefore falls within the scope of the prohibition, in Article 33(1) of the Sixth Directive, of other national taxes which can be characterised as turnover taxes, provided that, for a representative sample of businesses subject to both taxes, the ratio between the amounts paid in VAT and the amounts paid in the disputed tax is substantially constant.

129. That condition must be assessed by the national court, having regard to the detailed characteristics of the tax in issue.

II - Temporal limitation of the effects of a judgment given by way of preliminary ruling

A - General considerations

130. The last of the three questions posed by the Court before the second hearing was prompted by Advocate General Jacobs's Opinion and is that which has attracted the greatest response from the Member States. It concerns the circumstances and manner in which the effects of a preliminary ruling delivered by the Court may be subjected to a temporal limitation.

131. In limiting the temporal effect of a judgment a balance must always be sought between, on the one hand, the principle of correct and consistent interpretation of the law and, on the other hand, the principle of certainty in legal relationships conducted in good faith, even on the basis of an erroneous interpretation.

1. Validity of Community measures and temporal effect

132. The Court's practice has its roots in the second paragraph of Article 231 EC, which provides, with regard to actions for the annulment of a Community measure:

"If the action is well founded, the Court of Justice shall declare the act concerned to be void.
In the case of a regulation, however, the Court of Justice shall, if it considers this necessary, state which of the effects of the regulation which it has declared void shall be considered as definitive."

133. On the direct basis of that provision, the Court has adopted essentially two types of approach, by either declaring the effects of the contested measure to be definitive, (76) or by maintaining those effects until a new, valid measure has been adopted - sometimes specifying that this must be within a reasonable period. (77)

134. But the Court may also review the validity of a Community measure in response to a request for a preliminary ruling. Essentially in order to maintain consistency between the two situations, it has therefore applied the second paragraph of Article 231 EC by analogy in appropriate cases of that sort. Here, it has taken various approaches. In some cases, it has specified that the finding of invalidity has no retroactive effect whatever. (78) In others, it has stated that the effect is in general not retroactive but has allowed retroactive effect for those who have before the date of the judgment brought proceedings based on the invalidity. (79) In one case it has maintained the effects of invalid rules until new rules were adopted, (80) and in one very specific set of circumstances, when a provision was found to be invalid not for what it contained but for what it omitted, the Court stated simply that it was for the institutions concerned to adopt the measures necessary to correct the incompatibility. (81) Indeed, the choice of approach has in all instances been determined by the specific circumstances of the case.

2. Interpretation of Community law and temporal effect

135. The Court has also in certain cases (82) imposed a temporal limitation on the effects of a preliminary ruling on the interpretation of a provision of Community law.

136. Most of those cases fall into two broad categories: discriminatory failure to accord a pecuniary benefit, and imposition of a national tax contrary to Community law. Both categories may involve very large amounts of money wrongly withheld or levied, payment or reimbursement of which might cause serious difficulties for the finances of the Member State concerned. (83)

a) Effect ex tunc

137. In such cases, after reaching an interpretation as a result of which the national conduct in question must be considered unlawful, the Court has systematically recalled that its rulings on interpretation define the meaning and scope of a Community rule as it ought to have been understood and applied from its entry into force.

b) Effect ex nunc

138. Exceptionally, where the Member State had good grounds for believing that its conduct was compatible with Community law and there was a risk of serious economic repercussions, the Court has gone on to decide that its interpretation may not be relied upon to call in question situations established in good faith in the past: the interpretation takes effect from the date of the judgment itself. (84)

c) Exceptions to limitation of temporal effect

139. To date, the Court has excluded from any temporal limitation claims relying on the interpretation given but raised prior to the date of its judgment. (85)

d) The proposal in Meilicke

140. In Meilicke, (86) a case currently before the Court, Advocate General Tizzano has, in his Opinion of 10 November 2005, proposed a different approach from that hitherto followed by the Court.

141. He concluded first that a German rule on the taxation of dividends was contrary to Community law ­- as was clear from the Court's earlier judgment in Verkooijen (87) concerning a comparable rule of Netherlands law - but that the criteria for limiting the temporal effect of the future judgment were satisfied.

142. He then considered (88) what form the limitation should take, examining a broader range of possibilities than the Court has hitherto made use of. His conclusion was that the ruling should have effect from the date of the judgment in Verkooijen, when the correct interpretation of Community law had become clear. (89) There should none the less be an exception from that limitation for claims raised before that date. Moreover, it would be unjust to refuse all claims raised after the ruling in Verkooijen, although the publicity surrounding the Meilicke case had led to so many claims that an exception for all of them might well fail to prevent the risk of serious economic repercussions.

143. He consequently proposed that the exception should apply to all claims raised prior to the date on which notice of the order for reference in Meilicke was published in the Official Journal of the European Union, when it could be supposed that the possibility of a refund attracted the attention even of less diligent claimants.

e) Effect from a future date

144. In order to complete the range of possibilities that may be envisaged, it may be helpful to examine another option available in the (constitutional) law of some Member States (90) in situations resembling that of the present case, (91) namely that of setting a future date from which the judgment may be relied on.

145. In those systems, the aim appears to be above all to avoid situations in which no law would be worse than a bad law (including for example those in which the State would no longer be able to collect certain revenue or make certain payments) and to allow the legislature sufficient time to enact a measure which meets the requirements of the higher norm. In many cases, the court's competence to fix a future date is limited by law to a specific maximum period. Within that period, the choice may be determined by the court's assessment of a reasonable time for enacting new legislation, or by the periodicity of the legislation involved (in the case, for example, of tax legislation). (92)

3. Specific nature of preliminary rulings on interpretation of Community law entailing possible invalidity of national measures

146. Before going further, it may be helpful to draw attention to specific features of a preliminary ruling of this Court interpreting Community law in such a way as to entail the invalidity of a national measure, which may condition the options available to the Court when deciding to limit the temporal effects of its ruling.

147. The Court does not have jurisdiction to rule on the compatibility of a national measure with Community law, but only to interpret the latter in such a way as to enable the national court to determine such compatibility. (93)

148. In the present case the Court may of course interpret the Sixth Directive in such a way that the referring court will find that IRAP is not compatible with Community law. That is however very different from a declaration that the national tax in issue is itself invalid, a declaration which can only be made by the competent national court where appropriate, with effect from a date determined by that court or by operation of national law, and contrasts with the situation in a preliminary reference on the validity of a Community measure, which the Court can, in effect, annul directly.

149. The date from which the Court's ruling takes effect will be relevant when determining whether a litigant may or may not rely in a national court on the incompatibility of the national measure with Community law which is apparent as a result of the ruling, in respect of a time at which that measure was applied. It will not however be the only relevant factor, since national procedural rules may limit the extent of a claim in other ways.

B - Limitation of temporal effect in the present case

150. Even if the general considerations which I have set out above may be of assistance to the Court, any decision to limit the temporal effects of one of its judgments must be taken on a case-by-case basis, in the light of each set of circumstances.

151. Three questions must therefore be asked in the present case: Are there grounds for limiting the temporal effect of the judgment? If so, from what date should it be possible to rely on the ruling? And should any exception be made in favour of claims raised before a certain date?

152. In what follows I shall assume, in order to avoid constant repetition of the necessary proviso, that the Court will provide an interpretation as a result of which IRAP will prove incompatible with Community law.

1. Grounds for a limitation of temporal effect

153. Two essential criteria must be met, namely that those concerned should have acted in good faith and that there should be a risk of serious difficulties in the absence of a limitation. (94) Advocate General Jacobs considered that both criteria were met in the present case.

154. All the Member States which have commented on this aspect agree that the criteria are met here. Many of them have also submitted that the Court should assess those criteria more generously than it sometimes has in the past, referring, inter alia, to paragraph 42 of Advocate General Tizzano's Opinion in Meilicke. The Commission also agrees that the criteria are met, and Banca Popolare does not object in principle to the fixing of a temporal limitation.

155. I concur fully with Advocate General Jacobs's analysis on this point. On the undisputed evidence put to the Court, only a stricter approach to temporal limitation than that which has been followed in the past would justify a decision that there should be no temporal limitation in the present case. No argument has been made for a change of that kind, nor does there appear to be any reason in the circumstances of the present case to undertake a review of past practice. That being so, there is no need to consider whether a more generous approach to the question of temporal limitation may be called for in any future cases.

156. I would however stress that the circumstances of this case are very particular. In 1997 the Italian Government received from the Commission what could only be interpreted as an explicit assurance that IRAP would be compatible with Community law. (95) The amount of tax which may be claimed back has been stated by the Italian Government to be some EUR 120 billion, and that figure has not been contested. (96) The proceedings have been protracted, both at the national level and before the Court. (97) And in the latter context, the Grand Chamber has taken the rare step of reopening the oral procedure and holding a second hearing, a step which has attracted even more attention.

2. Choice of date for temporal limitation

157. In the past, where the Court has specified a temporal limitation subject to an exception for claims raised before a certain date, it has always used the same date as the basis for both the limitation and the exception: the limitation does not apply to claims already raised before the date from which the judgment may be relied on (in practice, hitherto always the date of its delivery). However, as Advocate General Tizzano has proposed in his Opinion in Meilicke, two different dates may be more appropriate in some circumstances. I shall deal in this section only with the principal date, that of the temporal limitation itself.

158. Clearly, if the temporal effect of the judgment is to be limited, it must be either from the date of delivery of the judgment itself or from some other specified date before or after that delivery.

159. It does not seem to me possible in the present case to identify any specific date in the past from which the judgment should take effect. In the circumstances of Meilicke, Advocate General Tizzano pointed out that the scope of the relevant provisions of Community law had become clear from the delivery of the judgment in Verkooijen. There does not appear to be any equivalent date here. A parallel with Dansk Denkavit is of course possible, but aspects of the case-law subsequent to that judgment, together with the conflicting views expressed in the present case, which have given rise to a second hearing, make it more difficult to regard Dansk Denkavit as a statement of the law equivalent to that in Verkooijen. The question whether a national tax such as IRAP, as described by the referring court, is incompatible with the Sixth Directive will not be authoritatively answered until the Court has delivered judgment in this case - which may perhaps be the "Verkooijen" for future cases.

160. The date of that delivery might therefore appear to be an appropriate date and, if it were chosen, would follow the pattern of this Court's consistent precedent.

161. Nevertheless, having regard to the specific circumstances of this case, it would also be possible to take a different approach and fix a future date, inspired by the practice of those national courts which, in similar circumstances, allow the legislature a reasonable period within which to enact a new, compatible measure, and by the Court's own practice, with regard to the validity of Community measures, in the case-law cited in footnotes 77 and 80.

162. Such an approach would in my view be more appropriate here. Realistically, the Italian authorities cannot be expected to change overnight their whole system of financing regional expenditure, nor can they be expected to have changed it in anticipation of the Court's judgment. If all taxpayers could rely immediately on that judgment in order to reclaim sums levied in IRAP as from the date of its delivery, that would be tantamount to abolishing the tax, and the means of financing the Italian regions, with immediate effect.

163. On the other hand, a date should not be set too far in the future. If it would be unreasonable to expect the immediate replacement of one tax by another, it is not unreasonable to suppose that the Italian authorities have by now made contingency plans for such replacement. They learned of the Commission's view that IRAP was incompatible in March 2004, when that institution lodged its observations to the Court, although those observations were not public, and they learned of Advocate General Jacobs's view that it was incompatible in March 2005, when he delivered his Opinion publicly in open court. Indeed, it seems that the gradual elimination of IRAP has been planned in the legislation since April 2003, and may be presumed to have progressed since then. (98)

164. In order to ensure that IRAP is replaced by a financing mechanism compatible with Community law in as smooth a manner and as short a time as possible, the appropriate date would appear to me to be the end of the tax period current when the Court's judgment is delivered. (99) Since the judgment will presumably be delivered this year and the tax period for IRAP appears to be a calendar year, that would mean that the ruling could not be relied upon to bring proceedings concerning IRAP levied in respect of any period of assessment ending on or before 31 December 2006 but that, in the event of delay in reforming the system, any sums levied in respect of any subsequent period could be reclaimed.

3. Exceptions to the date for temporal limitation

165. The limitation which I propose seeks to take account of the Member State's legitimate interests in a specific set of circumstances. It runs however counter to the principle that individuals are entitled to obtain repayment of national charges levied in breach of Community provisions. (100) Care should therefore be taken to ensure that the encroachment on that principle is not excessive, again having regard to the specific circumstances of the case.

166. Commonly, in the practice both of this Court and of the courts of a number of Member States, an exception from a limitation of the temporal effect of a judgment is made in favour of claims raised before a certain date. It may cover all those raised before the date of the judgment finding the incompatibility (as has been the usual practice of this Court), or only some of them; it may also be possible to choose a different cut-off date (as proposed by Advocate General Tizzano in Meilicke).

167. The notable difficulty in the present case is the apparently vast number of claims that have been raised for repayment of IRAP on the ground of its presumed incompatibility with Community law. Since these proceedings have attracted a great deal of attention over a relatively long period, it seems plausible that many of those claims, particularly the more recent among them, are of a speculative nature, introduced at no great effort or expense with a view to profiting from the forthcoming judgment. An exception in favour of such claims not only seems less justified because of their nature, but would also be likely, because of their number, to undermine seriously the effect sought by the limitation.

168. If an exception is to be made for claims brought before a particular date and if that date is to be chosen in the light of the considerations outlined in the previous paragraph, it should meet the following conditions: it should be as objective as possible; it should be likely to distinguish as far as possible between earlier claims brought in the belief that they were well founded, but none the less at some risk in view of the uncertainty of their outcome, and those brought at a later date in the light of a perceived probability of success; and it should be such as to dispose substantially of the problem posed by the extremely large number of claims.

169. Following Advocate General Tizzano's suggestion in Meilicke, Banca Popolare, the Italian Government and the Commission, together with several other Member States, have all proposed the date of 21 January 2004, when notice of the Court's receipt of the order for reference in the present case was published in the Official Journal. (101)

170. It seems to me however that if reasoning comparable to that underlying Advocate General Tizzano's proposal is applied to the present case, it should lead to a different date. His ground for proposing the date of the publication in the Official Journal in Meilicke was that it could reasonably be supposed to be the date on which the attention even of less diligent claimants was attracted to the possibility of a refund. (102) The interpretation of Community law had been clear since the judgment in Verkooijen, but the likelihood of its application to the German tax rule did not become real until a reference was made to this Court. Once potential claimants were informed of that reference, it seemed plausible that proceedings could be brought with a reasonable chance of success.

171. I do not think that publication in the Official Journal provided equivalent information in the present case, since the "Verkooijen element" is missing. (103) The request for a preliminary ruling raised the question of the compatibility of IRAP with Community law, but the result of the reference was by no means as apparent as in Meilicke, where the relevant provision of Community law had already been interpreted by the Court in parallel circumstances. Here, as I have shown, the case-law has not been entirely unequivocal in its approach to compatibility with the Sixth Directive.

172. However, it appears uncontested that there is now a widespread perception - or at least hope - in Italy that the Court is likely to deliver a judgment as a result of which IRAP will be found incompatible with Community law. The identification of a date from which that perception became concrete is bound to be to some extent arbitrary, but in my view the delivery of Advocate General Jacobs's Opinion on 17 March 2005 is the least arbitrary and most objective date in that regard. (104) It was from that point that a real likelihood emerged that the Court would give such a ruling; by contrast, if his Opinion had reached a different conclusion, the likelihood would have been considerably less.

173. I am therefore of the opinion that an exception to the temporal limitation of the judgment, in favour of those who brought proceedings before 17 March 2005, is likely to distinguish between earlier claims and those brought only when the probability of success was perceived to be stronger. And although an earlier date is proposed by the Italian Government, it may still be presumed that the risk of undermining the effects of the limitation will not be unduly exacerbated.

174. There remains however a concern raised in particular by the Netherlands Government.

175. As Advocate General Jacobs noted at point 85 of his Opinion, any shortfall in revenue caused by reimbursement of IRAP will have to be made up by other taxation. The Netherlands Government is concerned that injustice might ensue if all taxpayers (including new market entrants who had not been subject to IRAP) must make good the shortfall while only some of them benefit from the reimbursement. Moreover, it reasons, if the burden of the tax has been passed on to consumers (as must in general be the case if IRAP is caught by the prohibition in the Sixth Directive), the benefit of the reimbursement will be multiplied.

176. That reasoning might appear to suggest that in fact no exceptions should be made to the limitation of the effects of the ruling. I accept that there may be circumstances in which that is the case (and that in such circumstances it may be inappropriate to make an exception even for the claimants in the proceedings giving rise to the reference, since the identity of the case in which a reference is made is unrelated to any diligence on the part of the claimant in bringing the proceedings), but I do not think that they apply here.

177. The problem of unequal treatment as between taxpayers will not be significant if the number of claimants benefiting from the exception is sufficiently circumscribed - which seems likely to be the case if a date even later than that proposed by the Italian Government is adopted. And the problem of unjust enrichment through reimbursement of a tax the burden of which has been passed on has been adequately dealt with in the Court's case-law, most recently in Weber's Wine World, (105) a sequel to EKW.

C - Consequences for other Member States

178. I turn finally to consider briefly the wider consequences of the ruling (which I still presume for present purposes to entail the incompatibility of a tax fitting the description of IRAP given by the referring court) and of any limitation of its temporal effect.

179. If a temporal limitation is imposed on the effect of such a ruling, it will be in the interest of the Member State concerned, in order to avoid exceptional disruption. If there is an exception to the limitation it will be granted, by contrast, in the interests of those who, within the Member State, have sought to assert claims in reliance on Community law.

180. Yet a ruling on interpretation has general effect. If the Court rules that a tax having the characteristics of IRAP as described by the referring court is incompatible with the Sixth Directive, that will be true for IRAP and equally true for any other tax having those characteristics in any other Member State.

181. However, any temporal limitation and any exception thereto decided upon by the Court will be based on an assessment of the situation - existence of good faith on the part of the State, risk of serious disruption for the State and need for effective judicial protection of diligent claimants - in Italy, and that assessment might be quite different with regard to another Member State which also applied a tax having the same characteristics.

182. That consideration implies that any limitation should be not only temporal but also, in effect, spatial - a point of some relevance in the present case since it appears from several of the numerous articles which have already appeared in legal and tax journals concerning this case that one or more Member States other than Italy may apply taxes which, at least in the opinion of some authors, share certain characteristics with IRAP.

183. It is not of course possible for the Court to decide in the present case whether a limitation of temporal effect would be appropriate with regard to such other taxes or, if appropriate, what date should apply and what exceptions, if any, should be made. Yet the Court has consistently held that a limitation of temporal effect may be allowed only in the actual judgment ruling upon the interpretation sought, (106) and the decision is particular to the factual context of the preliminary reference.

184. Consequently, in accordance with the basic principle governing the effects of all of the Court's preliminary rulings on interpretation, the judgment will apply ex tunc in regard to any other tax having the relevant characteristics in another Member State.

185. The difficulties which might arise as a result are similar to those outlined by Advocate General Tizzano in his Opinion in Meilicke, in particular at point 47 et seq. A scenario such as the following might be imagined.

186. If taxpayers challenge a national tax in reliance on the judgment in the present case, it will be open to the national court to refer a question for a preliminary ruling; indeed, there is no certainty that another national tax will possess the characteristics defined in this case so identically that all doubt as to compatibility is excluded. In the context of the proceedings on that reference, it will be open to the Member State concerned to seek a limitation of the temporal effects of the ruling to be given. If the circumstances are appropriate, the Court could decide on a limitation based on the date of the judgment in the present case - in a manner comparable to Advocate General Tizzano's proposal in Meilicke that the effects should be limited to the date of the judgment in Verkooijen - and on any exceptions that might be called for to the effects of the limitation.

III - Conclusion

187. In the light of all the foregoing considerations, I am of the opinion that the Court should rule as follows in response to the question raised by the Commissione Tributaria Provinciale di Cremona:

(1) a tax having the features of IRAP as described in the order for reference, that is to say which

- is levied on all natural and legal persons who regularly carry on an activity with the object of producing or trading in goods or providing services,

- is imposed on the difference between the proceeds and costs of the taxable activity,

- is charged in respect of each stage in the production and distribution process corresponding to a supply or set of supplies of goods or services made by a taxable person, and

- imposes a burden at each of those stages which is globally proportional to the price at which the goods or services are supplied,

falls within the scope of the prohibition, in Article 33(1) of the Sixth Council Directive 77/388/EEC, of other national taxes which can be characterised as turnover taxes, provided that, for a representative sample of businesses subject to both taxes, the ratio between the amounts paid in VAT and the amounts paid in the disputed tax is substantially constant.

That condition must be assessed by the national court, having regard to the detailed characteristics of the tax in issue.

(2) The prohibition in that article may not be relied upon in order to claim reimbursement of IRAP levied in respect of any period of assessment prior to the Court's judgment, or in respect of the period during which that judgment is delivered, except by persons who initiated legal proceedings or raised an equivalent administrative claim before 17 March 2005, the date on which Advocate General Jacobs delivered his Opinion in the present case. Such claimants may rely on it to the extent that their claims are not otherwise barred by national procedural rules which observe the principles of equivalence and effectiveness.


1 - Original language: English.


2 - Sixth Council Directive 77/388/EEC of 17 May 1977 on the harmonisation of the laws of the Member States relating to turnover taxes - Common system of value added tax: uniform basis of assessment (OJ 1977 L 145, p. 1, hereinafter "the Sixth Directive").


3 - I refer to that Opinion for the legislative, factual and procedural background to the case, which I shall reiterate or complete only when appropriate.


4 - I have simplified the wording of the questions as they appear in the order reopening the oral procedure.


5 - Point 27 of the Opinion. See also, for example, Case 252/86 Bergandi [1988] ECR 1343, paragraph 13; Joined Cases C-370/95 to C-372/95 Careda and Others [1997] ECR I-3721, paragraphs 25 and 26.


6 - Of Legislative Decree No 446 of 15 December 1997 introducing IRAP, published in GURI No 298 of 23 December 1997.


7 - For "valore aggiunto prodotto", namely "added value produced" in Italian.


8 - Opinion of Advocate General Jacobs, points 36 and 67.


9 - Since that provision is couched in permissive terms ("... this Directive shall not prevent a Member State from maintaining or introducing ... any taxes, duties or charges which cannot be characterised as turnover taxes ...") it might be thought preferable to view the prohibition as flowing rather from the directive as a whole, with its rather detailed harmonisation requirements, and from the second paragraph of Article 10 EC, which prohibits any measure which could jeopardise the attainment of the objectives of the Treaty, with Article 33(1) of the directive clarifying the prohibition by specifying that it concerns only levies which can be characterised as turnover taxes. However, the difference of viewpoint is of no practical consequence, and I shall assume, with the consistent case-law, that the prohibition resides in Article 33(1).


10 - See points 2 and 8 above, and 22 below.


11 - A number of other cases are of lesser importance: Case 73/85 Kerrutt [1986] ECR 2219 concerned a form of stamp duty explicitly permitted by Article 33 of the Sixth Directive; in Joined Cases 317/86, 48/87, 49/87, 285/87 and 363/87 to 367/87 Lambert and Others [1989] ECR 787 (summary publication), the Court repeated the same ruling as in Bergandi (cited in footnote 5), with regard to the same tax; the tax in issue in Dansk Denkavit (cited in footnote 16) was also subsequently the subject of infringement proceedings in Case C-234/91 Commission v Denmark [1993] ECR I-6273, in which the Court followed the same reasoning; and in Case C-28/96 Fricarnes [1997] ECR I-4939, paragraph 34 et seq., the Court followed the same reasoning as in UCAL (cited in footnote 20), decided on the same day, with regard to a similar levy.


12 - Case 295/84 [1985] ECR 3759, especially paragraphs 14 to 17.


13 - Cited above in footnote 5, paragraphs 6 to 20.


14 - Joined Cases 93/88 and 94/88 [1989] ECR 2671, paragraphs 6 to 21.


15 - Case C-109/90 [1991] ECR I-1385, especially paragraph 14.


16 - Case C-200/90 [1992] ECR I-2217.


17 - Case C-347/90 [1992] ECR I-2947.


18 - Case C-208/91 [1992] ECR I-6709.


19 - Cited above in footnote 5.


20 - Case C-347/95 [1997] ECR I-4911, paragraph 30 et seq.


21 - Case C-130/96 [1997] ECR I-5053.


22 - Case C-318/96 [1998] ECR I-785.


23 - Joined Cases C-338/97, C-344/97 and C-390/97 [1999] ECR I-3319.


24 - Case C-437/97 [2000] ECR I-1157, especially paragraphs 19 to 25.


25 - Case C-101/00 [2002] ECR I-7487, especially paragraphs 91 to 107.


26 - Case C-308/01 [2004] ECR I-4777, especially paragraphs 23 to 37.


27 - Rousseau Wilmot, paragraph 13; Wisselink, paragraph 8; see also the preamble to First Council Directive 67/227/EEC of 11 April 1967 on the harmonisation of legislation of Member States concerning turnover taxes (OJ, English Special Edition 1967, p. 14, hereinafter "the First Directive"), eighth recital.


28 - Bergandi, paragraph 7; SPAR, paragraph 17; Pelzl, paragraph 14; see also the preamble to the First Directive, first and second recitals.


29 - Bergandi, paragraph 9; SPAR, paragraph 19; Pelzl, paragraph 18.


30 - Rousseau Wilmot, paragraph 16; Bergandi, paragraph 14; Wisselink, paragraph 17; Dansk Denkavit, paragraph 11; Bozzi, paragraph 9; Beaulande, paragraph 12; Careda, paragraphs 13 and 24; UCAL; paragraph 33; Solisnor, paragraph 13; EKW, paragraph 20.


31 - Paragraph 19.


32 - Rousseau Wilmot, paragraph 15; Bergandi, paragraphs 8 and 15; Wisselink, paragraph 18; Giant, paragraph 12; SPAR, paragraph 23; Pelzl, paragraph 16.


33 - That summary is based on Pelzl, paragraphs 20 and 21. The substance of the characteristics has remained constant, although there are slight differences in the precise formulation: see Dansk Denkavit, paragraph 11; Bozzi, paragraph 12; Beaulande, paragraph 14; Careda, paragraph 14; UCAL, paragraph 34; Solisnor, paragraph 14; EKW, paragraph 22; Tulliasiamies, paragraph 99; and GIL Insurance, paragraph 33.


34 - See Bergandi, paragraph 14; Giant, paragraph 11; SPAR, paragraph 22; Pelzl, paragraph 20.


35 - See point 22 above and the case-law cited there.


36 - Dansk Denkavit, paragraph 14; Careda, paragraph 14; Solisnor, paragraph 14; SPAR, paragraph 22; EKW, paragraph 21; GIL Insurance, paragraph 32.


37 - Paragraph 15 and ruling. It should be noted however that not all of those features were considered relevant in the Court's analysis; see point 28 and footnotes 50 and 52 to 54 below.


38 - At paragraph 15.


39 - At paragraphs 18 and 19.


40 - At paragraphs 24 and 25.


41 - Rousseau Wilmot, paragraph 16; Giant, paragraph 14; Pelzl, paragraphs 24 and 25.


42 - Bergandi, paragraph 16 et seq.; Wisselink, paragraph 20.


43 - Beaulande, paragraph 18; UCAL, paragraph 36; GIL Insurance, paragraph 36.


44 - Wisselink, paragraph 20; Bozzi, paragraph 16; Beaulande, paragraph 17; UCAL, paragraph 36; Pelzl, paragraph 23.


45 - Bozzi, paragraph 15; UCAL, paragraph 36; Tulliasiamies, paragraph 102.


46 - Bozzi, paragraph 16; Beaulande, paragraph 17; UCAL, paragraph 36; SPAR, paragraph 27; Tulliasiamies, paragraph 103; GIL Insurance, paragraph 36.


47 - At paragraphs 11 and 12.


48 - SPAR, paragraphs 25 and 26.


49 - Wisselink, paragraph 10; Careda, paragraph 17; Tulliasiamies, paragraph 98.


50 - Giant, paragraph 9; Dansk Denkavit, paragraph 15; SPAR, paragraph 21.


51 - Careda, paragraph 18.


52 - Dansk Denkavit, paragraph 15; Careda, paragraphs 23 and 25.


53 - Dansk Denkavit, paragraph 15.


54 - Ibid.


55 - Finland, France, Hungary, Italy and Spain.


56 - See Article 12(3)(a) of the Sixth Directive. The Commission has on several occasions proposed a maximum rate but the proposal has not been accepted, although a de facto maximum of 25% applies, perhaps for practical reasons related to the increased likelihood of fraud when tax rates rise.


57 - See also the Opinion of Advocate General Tesauro in Dansk Denkavit (point 8, sixth paragraph, at p. I-2235).


58 - At points 24 and 25 of his Opinion.


59 - Although it appears from paragraphs 11 and 12 of Wisselink that this conclusion is not absolute, and even a single-stage tax may in some cases prevent the proper working of the common VAT system.


60 - There is however a suggestion in Bergandi, at paragraph 17, that a fixed-rate tax based on an objective evaluation of foreseeable receipts might be caught if passed on to the consumer.


61 - In particular in Article 13(B)(d).


62 - Reports have been commissioned by the Commission on the application of VAT to financial services, including bank transactions; a 1996 study by Ernst & Young is available on the Commission's website.


63 - Compare points 39, 40 and 53 of Advocate General Jacobs's Opinion in the present case.


64 - See the case-law cited in point 16 above. The Court has of course taken account of the possibility of passing on the burden of the tax, which is a feature of indirect taxation, but only for the purposes of comparison with VAT, not in order to establish a categorisation as direct or indirect.


65 - Eighteenth Council Directive 89/465/EEC of 18 July 1989 on the harmonisation of the laws of the Member States relating to turnover taxes - Abolition of certain derogations provided for in Article 28(3) of the Sixth Directive (OJ 1989 L 226, p. 21).


66 - Thirteenth Council Directive 86/560/EEC of 17 November 1986 on the harmonisation of the laws of the Member States relating to turnover taxes - Arrangements for the refund of value added tax to taxable persons not established in Community territory (OJ 1986 L 326, p. 40); although the VAT directives are numbered sequentially, not all those in the sequence have been adopted in fact (or in sequential order).


67 - See point 27 and footnote 41 above. It may be noted that, unlike Dansk Denkavit, all of those judgments were delivered by a chamber of the Court.


68 - The French and Hungarian Governments have made similar submissions.


69 - See point 105 et seq. below.


70 - See in particular Pelzl, paragraph 27; see also, with regard to the compatibility of a locally-levied tax with the prohibition of charges having an effect equivalent to that of a customs duty on exports, Case C-72/03 Carbonati [2004] ECR I-8027, paragraph 27 et seq.


71 - See paragraph 27 above, footnotes 41 and 45.


72 - At paragraph 18 of the judgment.


73 - With the perhaps unfortunate and minor exception of exempt transactions, if they form inputs for the purpose of further onward supplies.


74 - The possible exceptions are Rousseau Wilmot and SPAR, in neither of which the Court explicitly referred to the four "essential characteristics".


75 - See point 96 above.


76 - In a number of judgments from Case 34/86 Council v Parliament [1986] ECR 2155, paragraph 48, to Case C-445/00 Austria v Council [2003] ECR 8549, paragraphs 103 to 106.


77 - In a number of judgments from Case 81/72 Commission v Council [1973] ECR 575 to, most recently, Case C-178/03 Commission v Parliament and Council [2006] ECR I-0000.


78 - See in particular the first three cases in which the Court used this possibility: Case 4/79 Providence Agricole de la Champagne [1980] ECR 2823, paragraphs 42 to 46, Case 109/79 Maïseries de Beauce [1980] ECR 2883, paragraphs 42 to 46, and Case 145/79 Roquette Frères [1980] ECR 2917, paragraphs 50 to 52.


79 - See for example Case 41/84 Pinna [1986] ECR 1, paragraphs 26 to 30; Case C-228/92 Roquette Frères [1994] ECR I-1445, paragraphs 17 to 30.


80 - Case 300/86 Van Landschoot [1988] ECR 3443, paragraphs 22 to 24.


81 - Joined Cases 117/76 and 16/77 Ruckdeschel [1977] ECR 1753, and Joined Cases 124/76 and 20/77 Moulins Pont-à-Mousson [1977] ECR 1795.


82 - From, first and most famously, Case 43/75 Defrenne II [1976] ECR 455, paragraphs 69 to 75, to, most recently for the time being, EKW, cited in footnote 24, paragraphs 57 to 60.


83 - For example Case C-262/88 Barber [1990] ECR I-1889, which also illustrates the fact that serious difficulties are not sufficient to justify a temporal limitation: after noting in paragraphs 40 and 41 the very large number of workers who might be entitled to claim equal treatment on the basis of the Treaty interpretation given, the Court went on in paragraphs 42 and 43 to verify that Member States had been reasonably entitled to consider that another interpretation was possible.


84 - At a national level, the normal temporal effects of a ruling of incompatibility are also either ex tunc (the general rule, for example, in the courts of Belgium, France, Germany, Ireland, Italy, Spain, Poland and Portugal) or ex nunc (the general rule, for example, in some, if not all, courts of Austria, the Czech Republic, Greece, Hungary, and Slovenia; in the case of at least Greece and Slovenia, the judgments of some other courts normally have ex tunc effects), with an apparent preponderance of ex tunc.


85 - At a national level, a similar practice may be followed, whether the basic effect is ex tunc or ex nunc. Generally, in the interests of legal certainty, ex tunc effects may be relied upon only in proceedings which have been or may still be brought within the relevant procedural time-limits, although in some jurisdictions the effect may be formally restricted to the instant case because the court has no competence to annul a legislative measure, only to disapply it (the general rule, for example, in the courts of Denmark, Finland, Luxembourg, Sweden and the United Kingdom). Where ex nunc effects are concerned, an exception is commonly made for proceedings which at the date of the judgment have been or may still be brought within the relevant procedural time-limits; for those which at that date have already been brought; or for those which actually gave rise to the judgment.


86 - Case C-292/04.


87 - Case C-35/98 [2000] ECR I-4071.


88 - At point 43 et seq.


89 - Cases in which national courts will annul a legislative measure with effect from a specific date in the past appear to be rare.


90 - They include Austria, Belgium, the Czech Republic, Germany, Hungary, Poland, Slovenia and Spain. In other systems, there are fundamental objections to this approach, also described as "prospective". As Lord Nicholls of Birkenhead explained recently in National Westminster Bank v Spectrum Plus [2005] UKHL 41: "The essence of the principled argument against prospective overruling is that in this country prospective overruling is outside the constitutional limits of the judicial function. It would amount to the judicial usurpation of the legislative function."


91 - Namely, where an existing legislative rule of general application is declared incompatible with a higher norm; it will rarely if ever be possible to find a wholly similar situation, in which the court finding the incompatibility refers the case back to another court for a ruling on validity.


92 - The period between the date of the judgment and the date from which it may be relied on is sometimes referred to as "transitional". It should be noted however that this is a misnomer; the contested national measure is invalid and must be replaced with effect from a specific date.


93 - See, for example, Case C-265/01 Pansard [2003] ECR I-683, paragraph 18; see also the Court's information note on references from national courts for a preliminary ruling (OJ 2005 C 143, p. 1), point 6.


94 - See also, most recently, Case C-402/03 Skov [2006] ECR I-0000, paragraph 51.


95 - The fact that here the grounds for the Italian Government's belief lie in the Commission's assurance is of course not decisive; good faith could also be based on misleadingly drafted legislation, or indeed obscure case-law.


96 - This compares with a figure of between EUR 5 and 13 billion in Germany in the Meilicke case; see point 35 of the Opinion.


97 - Banca Popolare originally brought its action in 2001 and the reference was made in late 2003.


98 - See Legge delega per la riforma del sistema fiscale statale (L. 7/4/03, n. 80) in GURI No 91 of 18 April 2003, in particular Articles 8 and 10(4), (5) and (7) thereof.


99 - There are precedents in (at least) the German Bundesverfassungsgericht and the Belgian Cour d'Arbitrage for the choice of a future date corresponding to the end of a relevant current period, such as calendar year, tax year or academic year.


100 - See for example Case C-147/01 Weber's Wine World [2003] ECR I-11365, paragraph 93 and the case-law cited there.


101 - OJ 2004 C 21, p. 16.


102 - See point 62 of the Opinion.


103 - See my remarks at point 159.


104 - The earlier statement of the Commission's views, whether in its written observations, which are not made public, or even at the hearing, might also have been considered, but that cannot in my view be compared to the first public indication, in written, reasoned and generally available form, of the views of a Member of the Court.


105 - Cited in footnote 100, paragraphs 93 to 102.


106 - See EKW, paragraph 57, and the case-law cited there.


BAILII: Copyright Policy | Disclaimers | Privacy Policy | Feedback | Donate to BAILII
URL: http://www.bailii.org/eu/cases/EUECJ/2006/C47503.html

#Banca popolare di Cremona (Taxation) [2006] EUECJ C-475/03 (14 March 2006)