Case C‑167/05
Tribunal de Justicia de la Unión Europea

Case C‑167/05

Fecha: 18-Jul-2007

OPINION OF ADVOCATE GENERAL

MENGOZZI

delivered on 18 July 20071(1)

Case C‑167/05

Commission of the European Communities

v

Kingdom of Sweden

(Alcoholic beverages – Different tax treatment of wine and beer)





1.The present case concerns an action which the Commission has brought against the Kingdom of Sweden under Article 226 EC, seeking a declaration that the Kingdom of Sweden has failed to fulfil its obligations under the second paragraph of Article 90 EC by applying a system of internal taxes under which beer, which is mainly produced in Sweden, is indirectly protected, as compared with wine, which is mainly imported from other Member States,.

I–The national legislation at issue

2.In Sweden, the rate of tax on alcohol(2) is determined on the basis of the actual alcoholic strength by volume (alcoholic content) of the beverages.

3.The case-file indicates that when the Kingdom of Sweden acceded to the European Union in 1995, beer with an alcoholic strength in excess of 3.5% vol. (known as strong beer)(3) was subject to a rate of tax largely equivalent to the rate applicable to wine with an alcoholic strength in the region of 11% vol.

4.On 1 January 1997, the Swedish Parliament decided to reduce the excise duty on beer. According to the Swedish Government’s statement of defence, that decision was justified by the need, following Sweden’s accession to the European Union, to combat the growing border trade in beer with Denmark, which had been prompted by the differing levels of taxation in force in the two countries.

5.In response to pressure from the Commission, which had voiced doubts concerning the compatibility with the second paragraph of Article 90 EC of the arrangements under the 1997 amendment for taxing beer and wine, the Swedish Parliament decided to reduce the excise duty on wine with effect from 1December 2001.

6.Based on the system currently in force pursuant to Articles 2 and 3 of the lagen (1994: 1564) om alkoholskatt,(4) the rates of tax on the production and consumption of beer and wine are set as follows.

7.Excise duty per litre of beer is levied at the rate of 1.47 Swedish Kroner (SEK) by volume of alcohol. Beers with an alcoholic content of no more than 2.8% vol. are exempt from that duty (Article 2 of the lagen (1994: 1564) om alkoholskatt).

8.Excise duty per litre of wine is levied on a flat-rate basis, which increases according to the alcoholic content of the beverage and is determined in accordance with the following scale:

SEK 7.58 if the alcoholic strength is higher than 2.25% vol. but lower than or equal to 4.5% vol.;

SEK 11.20 if the alcoholic strength is higher than 4.5% vol. but lower than or equal to 7% vol.;

SEK 15.41 if the alcoholic strength is higher than 7% vol. but lower than or equal to 8.5% vol.;

SEK 22.08 if the alcoholic strength is higher than 8.5% vol. but lower than or equal to 15% vol.;

SEK 45.17 if the alcoholic strength is higher than 15% vol. but lower than or equal to 18% vol.

9.Wines with an alcoholic content of 2.25% vol. or less are exempt from the duty (Article 2 of the lagen (1994: 1564) om alkoholskatt).

10.In Sweden, the retail selling of wine, strong beer and spirits is a statutory monopoly. A wholly State-owned company called Systembolaget Aktiebolag (‘Systembolaget’) is responsible for administering the monopoly.

II–The pre-litigation procedure

11.On 28 February 2000, in response to complaints that the Swedish arrangements for taxing wine and beer were discriminatory, the Commission sent the Kingdom of Sweden an initial letter of formal notice, in which it stated that, according to the information in its possession, those arrangements appeared to be incompatible with Article 90 EC, since the effect of making wine subject to a higher rate of tax than beer – both in absolute terms and as a percentage of the final price of the product – was to reinforce consumer habits and consolidate the advantage enjoyed by beer (a domestic product) to the detriment of the commercial potential of wine (a product imported from other Member States).

12.The Kingdom of Sweden replied by two letters, dated 10 May and 20November 2000 respectively, in which it pointed out that tax arrangements under which the excise duty on alcohol is determined according to the alcoholic strength by volume of the beverages is not in breach of Article 90 EC. However, it stated that it was prepared to reduce the rate of duty on wine by volume of alcohol, bringing it down to a level largely equivalent to the rate of duty on beer, and thereby restoring the taxation relationship that had existed between the two products prior to 1 January 1997.

13.On 19 June 2001, the Commission sent the Swedish Government a reasoned opinion, to which the latter replied by two letters, the first of 13 July 2001 and the second of 17 December 2001, informing the Commission that the rate of duty on wine had been reduced by 18.8% with effect from 1 December 2001.

14.On the view that the change which had been introduced did not neutralise the discriminatory nature of the Swedish tax arrangements, the Commission sent the Kingdom of Sweden a second letter of formal notice on 1 July 2002. The latter responded by letter of 24 September 2002.

15.On 9 July 2004, the Commission sent the Swedish Government a second reasoned opinion, calling upon it to adopt the measures necessary to comply with the reasoned opinion within two months.

16.After requesting an extension of the time‑limit, which the Commission granted, the Swedish Government replied by letter of 7 October 2004 contesting all the complaints levelled against it.

III–Procedure before the Court and the forms of order sought

17.By application lodged with the Registry of the Court on 14 April 2005, the Commission brought the present action.

18.By order of the President of the Court of 16 September 2005, the Republic of Latvia was granted leave to intervene in support of the forms of order sought by the Kingdom of Sweden.

19.Pursuant to Article 44(3) of the Rules of Procedure of the Court, the case was assigned to the Grand Chamber, following a request from Sweden to that effect by letter of 4 April 2006.

20.The Commission and the Kingdom of Sweden presented oral argument at the hearing on 8 May 2007.

21.The Commission claims that the Court should:

–declare that the Kingdom of Sweden has failed to fulfil its obligations under the second paragraph of Article 90 EC by applying a system of internal taxes under which beer, which is mainly produced in Sweden, is indirectly protected, as compared with wine, which is mainly imported from other Member States; and

–order the Kingdom of Sweden to pay the costs.

22.The Kingdom of Sweden contends that the Court should:

–dismiss the action; and

–order the Commission to pay the costs.

23.The Republic of Latvia contends that the Court should:

–dismiss the action.

IV–Legal analysis

A–Arguments of the parties

24.The Commission considers that the Swedish tax arrangements at issue are discriminatory and have a protectionist effect in relation to the trade in beer, which is largely a domestic product, to the detriment of the trade in wine, which is mainly imported from other Member States. Consequently, those rules are in breach of the second paragraph of Article 90 EC.

25.The Commission points out that the tax by volume of alcohol which is levied on a wine of medium alcoholic strength is approximately 20% higher than the tax on beer. The impact of that difference on the final price of the products is compounded as a result of the application of VAT at a rate of 25%, which is calculated on the basic price inclusive of excise duty.

26.According to the Commission, a competitive relationship exists between strong beer (with an alcoholic strength in excess of 3.5% vol.) and wine in the intermediate bracket, with an alcoholic strength of between 8.5% vol. and 15%vol.(5) and a price ranging between SEK49 and SEK70.(6) Since the selling of either product on a retail basis is a State monopoly, they will be competing with each other in the same sales outlets, belonging to the Systembolaget chain, or in specially licensed catering establishments. The Commission maintains that the fact that consumers have to go to Systembolaget to purchase the products in question increases their interchangeability.

27.The Commission considers that the tax arrangements for wine and beer restrain the potential consumption of wine and have the effect of protecting, indirectly at least, the trade in beer.

28.The statistical information on which the Commission relies(7) allegedly demonstrates that the rate of tax on the products in question influences consumer behaviour. In particular, those statistics are alleged to show that, in 1997 – the year in which the rate of duty on beer was reduced – sales of beer rose, while sales of wine declined. Similarly, in 2002, after the rate of duty on wine was cut, wine sales underwent the highest increase ever recorded between 1995 and 2004.

29.The Commission points out that the impact of the excise duty on the final price of wines with an alcoholic strength of between 8.5% vol. and 15% vol. is more significant within the SEK41 to SEK70 price bracket, that is to say, the product category which accounts for approximately 80% of total sales. According to the Commission, the impact is also significantly greater than the impact on the category of beer most frequently sold, whatever the reference quantity used for making an assessment.(8)

30.The Kingdom of Sweden does not dispute the Commission’s claims that, by and large, beer is a domestic product and wine an imported product; that different methods are used to determine the level of tax to which the two products are subject; or, finally, that the products are to some extent in competition with each other.(9) According to Sweden, however, the products are not completely interchangeable, and it is not clear which categories of wine and beer are in competition with each other within the meaning of the second paragraph of Article 90 EC.(10)

31.However, the Swedish Government, supported by the Republic of Latvia, considers that the fiscal provisions at issue are not intended to protect beer production indirectly from competition from wine, nor do they have such an effect, and, consequently, no breach of the second paragraph of Article 90 EC may be imputed to that Government.

32.According to the Swedish Government, it is clear from the case-law of the Court that a breach of that nature may be regarded as established only if it can be demonstrated that the tax at issue affects the competitive relationship between the products in question, by reducing potential consumption of the imported product to the benefit of the competing domestic product. It is essential to take into account, for that purpose, the difference between the selling prices of the products and the effect of the tax on those prices, in order to ascertain whether the difference in the taxation of the products is likely to influence consumer choices.

33.In the present case, if the average final price of a litre of strong beer in the category most frequently sold is compared with that of a litre of wine(11) of alcoholic strength in the region of 11% vol. in the most popular and least expensive category, it is clear – according to the Kingdom of Sweden – that the difference in the taxation of the two products is not likely to influence consumer behaviour. In fact, even if the wine were to be subject to the same taxes as are levied on beer, it would still cost approximately twice as much as beer. On the basis of that information, the Swedish and Latvian Governments consider that the Commission has failed to demonstrate that the effect of Sweden’s tax arrangements is to protect beer production from competition from wine. The Republic of Latvia also points out that the statistics show that the period from 1997 to 2004 saw a greater increase in the consumption of wine in Sweden than of beer.

34.The Swedish Government agrees with the Commission that Swedish consumers of alcohol are sensitive to variations in the price of the products. However, according to that Government, the fact that the consumer response to price cuts is to engage in increased consumption does not imply that consumer behaviour is also affected in terms of the choice consumers make between the different categories of alcoholic beverage. On that point, the Latvian Government notes that demand for the products in question is relatively impervious to price variations and that price itself does not appear to be a factor determining consumer choice.

35.The Swedish Government also contends that the Systembolaget sales statistics presented by the Commission are incomplete, since, on the one hand, they fail to take account of sales in catering establishments and, on the other, they do not make it possible to isolate the data relating to sales of wine in the category which is in competition with strong beer. In its statement of defence, the Swedish Government reproduced a table showing sales of the products at issue during the period from 1996 to 2004, and incorporating the data derived from the statistics of the Systembolaget, the Skatteverket (Swedish authority responsible for direct and indirect taxation) and the Svenska Bryggareföreningen (Swedish federation of beer, cider, soda-water and water producers).(12)

36.In conclusion, the Swedish Government, supported by the Latvian Government, argues that the Commission has failed to prove that the tax arrangements at issue have a protectionist effect and, consequently, has also failed to prove the breach of the second paragraph of Article 90 EC imputed to Sweden in these proceedings.

B–Analysis

37.Under the first paragraph of Article 90 EC:

‘No Member State shall impose, directly or indirectly, on the products of other Member States any internal taxation of any kind in excess of that imposed directly or indirectly on similar domestic products.’

38.Under the second paragraph of Article 90 EC:

‘Furthermore, no Member State shall impose on the products of other Member States any internal taxation of such a nature as to afford indirect protection to other products.’

39.It has consistently been held by the Court that the purpose of Article 90 EC, as a whole, is to ensure the free movement of goods between the Member States under normal conditions of competition, by eliminating all forms of protection which might result from the application of discriminatory internal taxation against products from other Member States, and to guarantee absolute neutrality of internal taxation as regards competition between domestic and imported products.(13) To that end, the first paragraph of Article 90 EC prohibits any tax legislation the effect of which is to impose, through any kind of fiscal measure, a higher charge on an imported product than on a similar domestic product, whereas, according to the Court, the main function of the prohibition laid down in the second paragraph of Article 90 EC is to prevent all forms of indirect fiscal protectionism in the case of imported products which, without being similar to domestic products within the meaning of the first paragraph of Article 90 EC, are nevertheless in competition, even partial, indirect or potential, with them.(14)

40.In order to determine whether particular national tax arrangements are compatible with the prohibition under Article 90 EC on discriminatory internal taxation, a complex assessment, comprising several stages, is necessary.

41.That assessment entails, above all, an investigation that will define the relationship that exists between the products – domestic and imported – covered by the arrangements at issue.

42.In the first paragraph of Article 90 EC, that relationship is expressed in terms of product similarity, whereas, for the purposes of applying the second paragraph, it is necessary – according to the clarification which the Court has provided – that a competitive relationship should exist, even if that relationship is only partial, indirect or potential, provided it is not merely fortuitous but lasting and characteristic.(15)

43.The Court has also made clear that in order to establish whether a competitive relationship exists within the meaning of the second paragraph of Article 90 EC, it is necessary to take into account ‘not only the present state of the market but also the possibilities for development within the context of free movement of goods at the Community level and the further potential for the substitution of products for one another which might be revealed by intensification of trade, so as fully to develop the complementary features of the economies of the Member States.’(16)

44.Applying those factors, the Court has acknowledged the existence of a certain measure of mutual substitutability between wine and beer, as a consequence – according to the Court – of the fact that they are capable, to a certain extent, of meeting identical needs.(17) That assessment, which the Court made in a case in which it was asked to rule on the compatibility of the arrangements in force in the United Kingdom for taxing wine, was confirmed in a subsequent judgment concerning the rates of VAT applied to wine and beer in Belgium.(18)

45.In confirming that a competitive relationship exists between wine and beer, the Court did not neglect to point out the significant differences between the two products in terms of their qualitative features (first and foremost, the difference in alcoholic strength), production methods (wine is essentially an agricultural product and, as such, subject to the vicissitudes of climate, whereas beer is largely an industrial product) and price structure. Those differences – as well as the fact that a great variety of wines exist, in terms of characteristics, quality and price – led the Court to consider that the competitive relationship between the beverages at issue that is material for the purposes of a fiscal comparison must be viewed as restricted, and to accept it solely for ‘those wines which are the most accessible to the public at large, that is to say, generally speaking, the lightest and least expensive varieties.’(19)

46.In the circumstances forming the subject-matter of this case, it is common ground that wine is mainly an imported product, whereas beer is mainly a domestic product.

47.Furthermore, the Swedish Government raises no argument against the Commission’s claim that the Swedish market does not display features which would preclude recognition of a competitive relationship – of the kind the Court has already accepted – between wines in the cheaper bracket and strong beer, that is to say, products which have the largest percentage shares of the relevant markets. Although there are some differences of opinion between the parties when it comes to identifying the category of wines which is actually in competition with beer on the Swedish market, both when identifying it on the basis of alcoholic strength by volume (8.5% vol. to 15% vol. in the Commission’s view and 11% vol. in the view of the Swedish Government) or on the basis of price range, there is none the less substantial agreement in principle that the products at issue are to some extent interchangeable.

48.The first condition for the applicability of the second paragraph of Article 90 EC is therefore satisfied.

49.In order to ascertain whether a specific national fiscal provision is compatible with Article 90 EC, it is necessary next to compare the tax arrangements to which the products at issue are subject, in order to identify any difference in the taxation applicable to them and to determine the extent of that difference.

50.If, as in the present case, the comparison is between products with characteristics that render them largely dissimilar, even though they may be in part interchangeable, it may, in certain cases, prove particularly difficult to identify the criteria that will facilitate a proper comparison.

51.In its judgment on the arrangements for taxing wine in the United Kingdom, the Court held that the most appropriate method for assessing the taxation relationship between beer and wine required a comparison of the tax burden by reference to three different factors: volume, alcoholic strength and product price.(20)

52.If that method is applied in the present case, the outcome is as follows.

53.If we compare the rates of taxation by reference to volume,(21) then given that the tax on a litre of beer of average alcoholic strength (5.2% vol.)(22) is SEK7.64(23) and the rate of tax on a litre of wine of alcoholic strength between 8.5%vol. and 15% vol. is SEK 22.08, then the rate at which wine is taxed by volume is approximately three(24) times higher than that of beer. In the present case, that outcome is not particularly significant, however, given that a difference in taxation by volume is normal under tax arrangements which use alcoholic strength as a tax base.

54.As regards the difference in taxation by reference to alcoholic strength, a wine with an alcoholic strength of 12.5% vol. is taxed by volume of alcohol at a rate some 20% higher than beer (SEK 1.77 as compared with SEK 1.47), and a wine with an alcoholic strength of 11% vol. at a rate some 36% higher (SEK 2 as compared with SEK 1.47). A still lighter wine, with an alcoholic strength of 10%vol., is taxed by volume of alcohol at a rate almost 50% higher than beer. It is scarcely necessary to point out that alcoholic strength is certainly the most pertinent assessment criterion in the present case, since it is used as the tax base for determining excise duty.

55.It is more difficult to make a comparative assessment of the effect of the tax on the final price of the products at issue. That difficulty basically arises as a result of the disagreement between the parties as to which prices should be used as a basis for calculation. According to the Commission, which is relying on the official Systembolaget statistics, the relevant price bracket – which covers some 79% of wine sales – ranges from SEK 41 to SEK 70 for a 75cl bottle (from SEK 55 to SEK 93 per litre). The average price to the consumer of a litre of wine with an alcoholic strength of 12.5% vol. is stated to be SEK 65.33, whereas the average price to the consumer of a litre of beer with an alcoholic strength of 5.2% vol. is SEK 30.33. However, the Swedish Government considers that it is more accurate to use as a basis an average selling price of SEK 69.58 for a litre of wine with an alcoholic strength of 12.5% vol.(25) If the information which the Commission has supplied is used as the basis for calculation, tax accounts for 33.8% of the final price; and the basic price of wine (SEK 30.18 per litre) increases by approximately 73% once excise duty has been added and by approximately 116% once VAT has been added. However, if the information supplied by the Swedish Government is used as a basis, then with a basic price of SEK 33.584 per litre for wine, those figures become 31.7%, 54.7% and 107% respectively. As regards beer, according to the information supplied by the Commission and not disputed by the Kingdom of Sweden,(26) tax accounts for approximately 25.2% of the final price; and the basic price of a litre of beer (SEK 16.62) increases by 45.9% once excise duty has been added and 82.4% once VAT has been added. Clearly, in the category of wines taxed at SEK 22.08 per litre (of between 8.5% vol. and 15% vol.), the impact of taxation on price structure is greater in percentage terms for less expensive wines of lower alcoholic strength.

56.Whatever criterion for comparison is used, the above figures show that the categories of wine that are in competition with beer are subject to a significantly higher rate of tax than beer.

57.In order to determine whether national tax arrangements are compatible with the prohibition under Article 90 EC, it is lastly necessary to show that the arrangements are discriminatory in nature.

58.In that regard, the Court has made clear that, so far as the first paragraph of Article 90 EC is concerned, discrimination results automatically from the difference in taxation, since, in the case of similar products which are broadly comparable, the prohibition laid down in that provision applies as soon as an imported product is taxed more heavily than a domestic product. However, as regards the second paragraph of Article 90 EC, given the difficulty in making a sufficiently precise comparison of the products at issue, that assessment must, according to the Court, be conducted on the basis of a more general criterion, that is to say, the indirectly protectionist nature of the fiscal provision at issue.(27)

59.That assessment is without doubt the trickiest assessment to be made under Article 90 EC and, in the context of infringement proceedings, may place on the Commission a burden of proof which it is hard to satisfy.

60.In an attempt to alleviate that burden, the Court has specified that, for the purposes of applying the second paragraph of Article 90 EC, it is sufficient to establish that a particular taxation system ‘is likely, in view of its inherent characteristics, to bring about the protective effect referred to by the Treaty.’(28) The Court went on to state that ‘[w]ithout … disregarding the importance of the criteria which may be derived from statistics from which the effects of a given taxation system may be measured, it is impossible to require the Commission to supply statistical data on the actual foundation of the protective effect of the tax system complained of.’(29)

61.In its judgment on the tax arrangements for wine and beer in Belgium, however, the Court rejected the Commission’s argument that once a competitive relationship is established between two products, any difference in the rates of tax applied to the same basis of assessment is contrary to the second paragraph of Article 90 EC, and explained that any assessment of the compatibility of a given tax with the second paragraph of Article 90 EC ‘must take account of the impact of that tax on the competitive relationship between the products [concerned].’(30) According to the Court, ‘[t]he essential question is therefore whether or not the tax is of such a kind as to have the effect, on the market in question, of reducing potential consumption of imported products to the advantage of competing domestic products.’(31) In the same judgment, the Court also held that ‘in considering to what extent a protective effect actually exists, the difference between the respective selling prices of beer and wine competing with beer cannot be disregarded.’

62.Relying on that precedent, the Kingdom of Sweden is maintaining, in the present case, that the Commission has failed to satisfy the burden of proof incumbent upon it, since it has failed to take account of the significant price difference between the two products at issue, which would have led it to rule out the possibility that the difference in the taxes applicable to those products exerts any influence on consumer choice.

63.As regards the central argument which the Swedish Government advances and which is the mainstay of its defence, I consider it appropriate, before taking my analysis any further, to consider whether that argument is well founded. To that end, it is necessary to call to mind briefly the context of the judgment on which the Kingdom of Sweden’s plea is based, as well as the conclusion which the Court reached on that occasion.

64.In the circumstances of the case which gave rise to that judgment, the Commission imputed to the Kingdom of Belgium the application of a rate of VAT on wine some 6% higher than the rate on beer. On conclusion of its analysis, the Court found that the difference between the rates of tax applied was not such as to influence consumer choice. In particular, the Court considered that the Commission had failed to demonstrate that the difference between the respective prices for the beer and wine in competition with each other was so small that the 6% difference between the VAT rates applied to the two products was in fact capable of influencing consumer behaviour and thus giving rise to a protectionist effect in favour of beer.(32)

65.A more detailed analysis of the facts of the case allows us to comprehend better the significance of the conclusion which the Court reached and to assess whether, as the Kingdom of Sweden suggests, it is appropriate to transpose to the present case the approach adopted in that earlier case.

66.The Belgian Government contended that the price of a litre of beer, inclusive of taxes, amounted to BEF 29.75 and that the price of a litre of wine of comparable quality was BEF 125. Since the Commission’s arguments challenging those facts were rejected in their entirety, it must be inferred that the Court based its analysis on those facts. The selling price to the consumer of wine in Belgium, inclusive of VAT, was therefore almost four times(33) that of beer, whereas the difference between the VAT rates applied to wine and beer respectively was, as we have seen, 6% (25% for wine and 19% for beer).

67.In his Opinion,(34) which the Court largely followed on this point, Advocate General Vilaça pointed out that if the VAT rate were standardised at 25%, thereby increasing the rate applied to beer by 6 percentage points, the difference in price between the two products would decrease by BEF 1.5 (from BEF 95.25 to BEF 93.75), whereas if the rate were standardised at 19% – that is to say, by reducing the rate applied to wine – the reduction would have amounted to BEF 6 (from BEF 95.25 to BEF 89.25). In neither case did the price relationship between wine and beer undergo significant change (-1.5% if the rate were standardised at 25%, and -6.3% if the rate were standardised at 19%). In the former case, which had to be regarded as the most likely in view of the statements issued by the Belgian Government in the course of the proceedings, the price of a litre of beer would have increased by 5%, whereas, in the latter case, the price of a litre of wine would have fallen by 4.8%.(35)

68.At this point, it is necessary to compare the information we have in the present case with the figures set out above, in order to assess whether, in the present case, it is possible to reach the same conclusions as the Court arrived at in Case 356/85 Commission v Belgium.

69.On the basis of the data provided by or unchallenged by the Kingdom of Sweden,(36) and applying to a litre of wine with an alcoholic strength of 12.5% vol. the excise duty applicable to beer (SEK 1.47 by volume of alcohol), the outcome is a price reduction in absolute terms of SEK 4.63 per litre (EUR 0.50),(37) and, in percentage terms, a reduction of 6.65%. The price ratio between wine and beer would decrease by 11.8%, although continuing to be of the order of 2 : 1.(38)

70.In the light of those figures, is it possible to conclude – as the Court did in Case 356/85, referred to above – that the difference in the prices of beer and wine in Sweden is so great that the 20% disparity(39) between the rates of duty applied to the two beverages is not capable of influencing consumer behaviour?

71.In my view, that question must elicit a negative response. A comparison of the data relating to the present case and the facts which gave rise to the judgment in Case 356/85 reveals that the two cases cannot, in fact, be regarded as comparable. While the Court was able to hold in Case 356/85 that, given a ratio of 1 : 4 between the price of beer and the price of wine, any reduction in the price of wine in the region of 5% did not appear capable of influencing consumer choice, the same cannot be claimed with certainty and without taking other factors into consideration if that ratio is 1 : 2, and any reduction in the price of wine is in the region of 6.5%.(40) In principle, the narrower the ratio between the prices of two competing products, the more interchangeable they become, and, consequently, the degree of price variation that could make them potential alternates shrinks.

72.On the basis of the foregoing, I consider it necessary to reject the argument of the Kingdom of Sweden to the effect that it is possible, merely by taking into account the difference in price between wine and beer, to rule out the possibility that the disparity between the excise duty rates applied to the two products is likely to have any influence on consumer choice.

73.Since I have reached that conclusion in the light of the information provided by the Swedish Government, I shall not dwell on the question – the subject of lengthy discussion between the parties in the course of the written procedure – of the appropriateness of the method of comparison which the Swedish Government adopted, and which consisted in comparing the price of a litre of wine with the price of a litre of beer. In that connection, I shall merely point out that the method proposed by the Commission, which involves comparing the prices of a glass of wine and a tankard of beer of different volumes, does not appear to be the most appropriate, particularly in view of the ways in which the two products are marketed. They are, in fact, mainly sold through Systembolaget’s sales outlets,(41) in circumstances in which it is definitely more straightforward to compare the prices of the beverages per litre, bearing in mind also that, as the Swedish Government pointed out in its observations, the average Swedish consumer tends to buy wine packaged in what is known as the ‘Bag in Box’ format.(42) In any event, the Commission has failed to provide any information to demonstrate that its method of comparison is more relevant – bearing in mind the habits of Swedish consumers and the characteristics of the market – than the method suggested by the Swedish Government.

74.At this point, it is necessary to assess whether the evidence provided by the Commission is sufficient to demonstrate that the tax arrangements at issue are likely to produce a protectionist effect in relation to beer and to the detriment of wine.

75.According to the Commission, it is clear from the Systembolaget’s statistics that consumer behaviour is influenced by the price variations caused by the differences in the rates of duty applied to the products at issue. The Commission points out that for 1997, the year in which the tax reform cutting the rate of duty on beer by some 40% entered into force, the statistics show both that sales of beer rose by 8.9%, reversing the trend of previous years, when they fell significantly, and that sales of wine fell by 3.7%. In 2002, the year in which the rate of duty on wine was cut by 18.8%, sales of wine rose by 11%. The Commission further points out that, while in the period between 1997 and 2001 sales of wine rose overall, sales of beer clearly increased more significantly in percentage terms. Finally, the Commission notes that, while, in 1995, sales of wine by volume were substantially equivalent to sales of beer (approximately 103 million litres of each beverage), the volume of beer sold in 2004 (approximately 173 million litres) outstripped by some 20% sales of wine (approximately 139 million litres). According to the Commission, that disparity, which occurred after the 1997 tax reform, must be ascribed to the difference between the rates of duty on the two beverages introduced under the reform.

76.The figures which the Commission has provided do not appear to me to be particularly meaningful, first, and above all, because they relate to sales of wine and beer generally, and do not therefore permit a separate analysis of sales trends in the competing categories. Secondly, while it is true that the statistics which the Commission has attached to its submission show that, from 1997 and until 2003, sales of beer continued to rise, that increase was not constant. In 1998, the year after the rate of duty was reduced, the increase in sales amounted to just 2.8%, which is the minimum figure for the period 1997-2003, whereas in 1999 sales rose by 15.4%, far higher than the increase in 1997, in response to the cut in tax. The years after 1999, with the exception of 2003, reveal a more or less constant increase in sales, in the region of 10%. In my view, that suggests that factors other than the tax rate influenced the rise in beer sales in the reference period and, consequently, the difference in the volume of wine and beer sales during that period. So far as wine is concerned, while the Systembolaget’s statistics record a fall in sales in 1997, the year in which the reduced rate of tax was applied to beer, sales of wine in fact rose again in the following year, increasing by 2.7%, almost the same as the rate of increase in beer sales recorded that year (2.8%). From 1998 onwards, sales of wine have increased, although not at a constant rate.

77.In the light of the above considerations, I am of the view that the figures submitted by the Commission prove at most that Swedish consumers are sensitive to the variations that result from changes to the tax arrangements applicable to the beverages at issue; and, reflecting local patterns of consumption, where there is traditionally a predilection for beer, that sensitivity is more marked in the case of wine.(43) Moreover, it is clear that the effect, whereby sales expand as a result of a cut in tax, is of limited duration and probably lasts for just one year following the entry into force of the change in the tax rules.

78.While it is possible to infer from the statistics produced by the Commission that demand for the products at issue is to some degree elastic depending on price – and the Kingdom of Sweden does not dispute that fact –(44) it is not, in my view, possible to extract from those statistics information concerning the impact of the price variations on the competitive relationship between the products that is sufficiently significant to enable us to establish that the increases in the sales of beer and wine respectively in 1997 and 2002 are attributable, at least in part, to a substitution effect. An impact of that nature depends on the level of competition between the products themselves, hence on the cross-price elasticity of demand; however, the Commission has provided nothing on which to base an assessment of that factor.

79.At this point, it is necessary to consider what constitutes the actual burden of proof which the Commission must discharge in infringement proceedings pursuant to the second paragraph of Article 90 EC.

80.I have already stated that the Court does not require the submission of statistical data or econometric studies which demonstrate that the tax arrangements at issue actually have a protectionist effect; nor could the Commission reasonably be required to furnish such evidence.(45) However, it must demonstrate that the tax system is likely to produce such an effect. In other words, it is incumbent on the Commission to assess all the objective features of the tax provision at issue in order to determine whether, in the light of the fiscal position that existed prior to its introduction and the characteristics of the market, there is evidence sufficient to raise a presumption that it is having an impact on the potential consumption of the products originating in other Member States and on the competitive relationship between the products in question.(46) If there are indicators which substantiate the Commission’s argument, then it is for the government concerned to justify the difference in tax treatment at issue(47) and/or to provide evidence to show that there is no breach of the second paragraph of Article 90 EC.(48)

81.Moreover, allocating the burden of proof in that way does not appear to run counter to the abovementioned judgment in Commission v Belgium, in which the data produced by the Belgium Government – and unsuccessfully challenged by the Commission – bore out the fact that the different VAT rates applied to wine and beer were largely neutral in terms of the competitive relationship between the products, invalidating the Commission’s argument concerning a possible protectionist effect of the tax system applicable to them.

82.In the present case, are there sufficient indicators to raise a presumption that the tax arrangements at issue are protectionist in nature?

83.Let us start by reviewing the features of those arrangements. As the Swedish Government has itself acknowledged, the 1997 tax reform, under which the rate of duty on beer was cut by 40%, altered the fiscal parity that had previously existed between strong beer and wine with an alcoholic content of approximately 11% vol.(49) Under the system currently in force, even though the rate of tax is determined for both beer and wine according to the alcoholic strength by volume, and, therefore, according to an objective criterion which reflects the characteristics of the products, the method of calculation differs.(50) The duty on beer is in fact determined by multiplying a fixed amount by the alcoholic strength by volume of the beverage, with the result that the taxation by volume of alcohol is the same whatever the alcoholic content of the beverage, whereas wine is subdivided into five categories according to its alcoholic content and, within each category, the rate of taxation by volume of alcohol varies from a maximum rate, for wines with a lower alcoholic content, to a minimum rate, for wines with a higher alcoholic content. In the fourth category, which encompasses the wines which are in competition with strong beer, the rate of taxation varies from SEK 2.45 by volume of alcohol for wines of alcoholic strength equivalent to 9% vol. to SEK 1.47 for wines of alcoholic strength equivalent 15% vol. The fiscal impact is, therefore, greater on the lighter wines, which are also the cheaper wines, which are more directly in competition with beer.

84.It is then necessary to evaluate the difference in the taxation of the products at issue as a result of applying the system described above. As I had occasion to point out earlier,(51) a comparison of the fiscal burden to which wine and beer are respectively subject shows that, whatever the criterion for comparison that is used (volume, alcoholic strength or product price), those categories of wine which are in competition with beer are subject to a significantly higher rate of excise duty than applies to beer.

85.As far as the characteristics of the market are concerned, the sensitivity of Swedish consumers to the price variations that have resulted from the changes to the tax arrangements applicable to the products at issue is largely undisputed. Furthermore, the comparison between the prices of the beverages in question, which has been made on the basis of the data provided by the Swedish Government, has not made it possible to rule out the possibility that the difference in the taxation of the beverages may affect competition.(52) And, of themselves, the statistics which the Swedish Government has supplied, indicating that, between 1996 and 2003, in percentage terms, sales of wine recorded a greater increase than sales of beer, do not constitute evidence that the tax system at issue has no protectionist effect.(53)

86.In those circumstances, I consider that there is sufficient evidence in the present case to conclude that the tax system at issue is likely to affect the potential consumption of wine, which is mainly imported from other Member States, and place that product at a disadvantage as compared with beer, which is essentially a domestic product, inasmuch as the latter ‘constitutes the most relevant reference criterion from the point of view of competition’.(54) The arguments advanced by the Kingdom of Sweden are not, in my view, sufficient to rebut that finding. I therefore consider that the action must be upheld.

V–Costs

87.Pursuant to Article 69(2) of the Rules of Procedure, the unsuccessful party is to be ordered to pay the costs if they have been applied for by the successful party. In the light of the conclusions that I have reached in relation to the unsuccessful arguments of the Kingdom of Sweden, and since the Commission applied for costs to be awarded against the Kingdom of Sweden, I consider that the Kingdom of Sweden should be ordered to pay the costs.

88.In accordance with Article 69(4), the Republic of Latvia, which has intervened in this case, should be ordered to bear its own costs.

VI–Conclusions

89.In the light of all of the foregoing, I propose that the Court should:

–declare that the Kingdom of Sweden has failed to fulfil its obligations under the second paragraph of Article 90 EC by applying to wines in competition with beer a rate of excise duty that is comparatively higher than the duty levied on beer;

–order the Kingdom of Sweden to pay the costs;

–order the Republic of Latvia to bear its own costs.


1 – Original language: Italian.


2– Under Article 3 of Chapter 1 of the Alkohollag (Law 1994: 1738 on alcohol), alcoholic beverage means a beverage with an alcoholic strength in excess of 2.25% vol. Alcoholic beverages are subdivided into spirits, wine, strong beer and beer.


3– According to the definition contained in Article 6 of Chapter 1 of the Alkohollag, referred to in footnote 2 above.


4– Law of 15 December 1994 concerning excise duties on alcoholic beverages.


5– In the calculations which it uses to substantiate its allegations, the Commission takes the example of wine with an alcoholic strength of 12.5% vol.


6– The information on the price range is mentioned specifically only in the reply. In its application, the Commission refers more generally to ‘cheap light wines’ or ‘cheap table wines’.


7– The Commission mainly uses Systembolaget’s sales statistics, which are published on the latter’s website.


8– In its application, the Commission compares the effect of the tax on the final price of a bottle (a 75cl bottle of wine and a 33cl bottle of beer), of one litre and a glass of the product (a 15cl glass of wine and 33cl glass of beer).


9– The Swedish Government considers that a degree of competition exists between the most popular strong beers sold and the most widely sold and least expensive wines with an alcoholic strength of approximately 11% vol.


10– In its statement of defence, the Kingdom of Sweden questions the intensity and the unambiguous nature of the conditions of competition between the two products.


11– In its reply, the Commission maintains that the most appropriate method of price comparison from the consumer’s point of view is to use as a reference quantity a glass of wine and a glass of beer, and it challenges the approach of the Swedish Government which is based on the price per litre. The Swedish Government, however, takes the view that the method adopted by the Commission fails to produce a clear result.


12– The Swedish Government included it its rejoinder the data unavailable to the public on which those statistics are based.


13– See, inter alia, Case 356/85 Commission v Belgium [1987] ECR 3299, paragraph 6, and Case 323/87 Commission v Italy [1989] ECR 2275, paragraph 7.


14 – See, to that effect, Case 27/67 Fink-Frucht [1968] ECR 298. See also, inter alia, Case C-168/78 Commission v France [1980] ECR 347, paragraph 6, and Case 356/85, cited above, paragraph 7.


15– Case 27/67, cited above. There is, however, a tendency in the case-law of the Court to refrain from attaching excessive relevance to the distinction between the concepts of ‘similarity’ and ‘competition’, and to ‘generalise’ the assessment to be made pursuant to Article 90 EC. In various judgments, the Court has refrained from classifying the products at issue as similar or in competition, and has merely acknowledged that they are at least potentially or indirectly in competition. See, for example, Case 168/78, cited above; Case 169/78 Commission v Italy [1980] ECR 385; and Case 171/78 Commission v Denmark [1980] ECR 447. For a more recent example, see Case C-230/89 Commission v Greece [1991] ECR I-1909.


16– Case 170/78 Commission v United Kingdom [1980] ECR 417, paragraph 6.


17– Case 170/78, cited above, paragraph 14.


18– The Court held that the characteristics of the Belgian market were not such as to call into question the assessment made in relation to the British market. See Case 356/85, cited above, paragraph 11.


19– Case 170/78 Commission v United Kingdom [1983] ECR 2265, paragraph 12. See also Case 356/85, cited above, paragraph 11, and Case C-166/98 Socridis [1999] ECR I-3791.


20– Case 170/78, cited above, paragraph 18.


21– Assuming identical volume for both beverages.


22– This is the alcoholic strength by volume to which both the Commission and the Swedish Government refer for beer.


23– Beer is taxed at the rate of SEK 1.47 per litre by volume of alcohol: 1.47 x 5.2 = 7.64.


24– 2.89.


25– That price is derived from the average of the prices of the 22 wines, purchased in bottles or wine boxes, which are most widely sold in the price bracket ranging from SEK 61 to SEK 70, which accounts for the bulk of sales.


26– Average price to the consumer of SEK 30.33 for a litre of beer with an alcoholic strength of 5.2% vol.


27– Case 170/78, cited above, paragraph 9, see also the earlier judgment in Case 27/67, cited above. However, the tendency within the case-law to ‘generalise’ the assessment under Article 90 EC, without making a clear distinction between the application of the first or second paragraph thereof (see footnote 15 above) means that the protectionist effect of the tax arrangements at issue is regarded as the main criterion for determining whether that provision has been infringed.


28– Case 170/78, cited above, paragraph 10. Emphasis added.


29– Case 170/78, cited above, paragraph 10.


30– Case 356/85, cited above, paragraph 15.


31– Case 356/85, cited above, paragraph 16. Emphasis added.


32– Case 356/85, cited above, paragraph 18.


33– 4.2 times. Considering the prices net of VAT (BEF 100 for wine and BEF 25 for beer), the differential was precisely 1 : 4.


34– Opinion delivered on 26 February 1987 in Case 356/85, cited above, [1987] ECR 3299, point 102 et seq.


35– See the Opinion of Advocate General Vilaça, cited above, point 109.


36– The comparison is based on the average price of a litre of wine with an alcoholic strength of 12.5% vol. It is scarcely necessary to point out that if lighter and less expensive wines – falling within the tax bracket relevant in this case – were taken into consideration, the price difference compared with beer would be less marked, whereas, as we saw above, the difference in taxation to the detriment of wine would be greater.


37– The calculation also includes the application of VAT at a rate of 25%.


38– It would decrease from 2.3 to 2.1.


39– Comparing the duty levied by volume of alcohol and using as a basis a wine with an alcoholic strength of 12.5% vol.


40– But this increases if less expensive wines of lower alcoholic strength are taken into consideration.


41– Sales in licensed catering establishments do not seem to account for more than 10% of total sales.


42– The ‘Bag in Box’ (wine box) system is a method of packaging liquids in which the liquid is first sealed in a bag which may be made of plastic or aluminium. The bag is then placed inside a cardboard box which is designed to make it easier to transport and store the product. The success of that kind of packaging is due to the fact that it is convenient and keeps the contents fresh by limiting contact with the air. Normally, in the case of wine, the box will have a capacity of 3.5 litres, 10 litres or 20 litres.


43– The statistics which the Commission has provided record a greater increase in wine sales in 2002 than in beer sales in 1997, none the less the reduction in the rate of tax on wine was approximately half of the reduction previously applied to beer.


44 – The increase in wine sales in 2002, following the cut in excise duty, is also apparent from the statistics provided by the Kingdom of Sweden. It is also clear from the Swedish Government’s written submissions that it is the more or less high but, in any event, appreciable elasticity in demand for alcoholic beverages in general that has enabled Sweden traditionally to use taxation to intervene in relation to prices and thereby to influence consumer habits. In the past, for example, raising the rate of tax on beverages with a high alcohol content, combined with other non-fiscal measures, enabled the Government to steer consumers of such beverages towards products with a lower alcohol content, mainly wine.


45– That proof may be particularly hard to provide if the products are only partly interchangeable.


46– Clearly, the more direct the competitive relationship between the products at issue, the less demanding the burden of proof that falls on the Commission, until the point is basically reached where the issue is a difference in the taxation of products which are practically the same.


47– See, for example, the case concerning the taxation of denatured alcohol in Italy: Case 46/80 Vinal [1981] ECR 77.


48– I must make it clear that this does not imply a reversal of the burden of proof, since it is in any event incumbent on the Commission to provide sufficient evidence to demonstrate prima facie the existence of a protectionist effect.


49– According to the Swedish Government.


50– The case-file does not indicate whether the method of calculation was the same for both wine and beer before the change came into force in 1997.


51– See points 52 to 56 above.


52– See points 63 to 73 above.


53– See Case 168/78, cited above.


54– See, to that effect, Case 170/78, cited above, paragraph27.

Vista, DOCUMENTO COMPLETO