1. 

Is there any such thing as unfair prices?

2. 

In the field of competition law, different jurisdictions have made different choices in respect of that question. For example, in a number of them, among which the United States, conduct of undertakings with market power which merely exploits customers is generally not considered to infringe that law. However, the choice made by the drafters of the EU Treaties is evidently different: point (a) of the second paragraph of Article 102 TFEU includes, as a type of prohibited abuse of dominant position, conduct which consists in ‘directly or indirectly imposing unfair purchase or selling prices or other unfair trading conditions’.

3. 

Nevertheless, in its practice, the Commission has been extremely reluctant to make use of that provision against (allegedly) high prices practiced by dominant undertakings. Rightly so, in my view. In particular, there is simply no need to apply that provision in a free and competitive market: with no barriers to entry, high prices should normally attract new entrants. The market would accordingly self-correct.

4. 

It may however be different in markets with legal barriers to entry or expansion and, in particular, in those in which there is a legal monopoly. Indeed, there may be markets which, because of their particular features, are not run efficiently when open to competition. Likewise, a government may have legitimate policy reasons to limit competition in a specific market, thereby sacrificing economic efficiency, in order to pursue other public aims.

5. 

That is precisely the case in the main proceedings.

6. 

The case at hand offers the Court an opportunity to clarify the conditions under which the imposition of high prices by a dominant undertaking might infringe point (a) of the second paragraph of Article 102 TFEU. In other words, the present case concerns prices set by dominant undertakings that may be abusive because, being excessively high, they exploit customers. Conversely, it does not concern prices which may be abusive because of their exclusionary effects on competitors.

7.

Article 13(4) of the Konkurences likums (Competition law) has the same wording as point (a) of the second paragraph of Article 102 TFEU.

8.

The applicant in the main proceedings, Autortiesību un komunicēšanās konsultāciju aģentūra – Latvijas Autoru apvienība (‘AKKA/LAA’), in its capacity as a collecting society, issues licences for the public performance of musical works in commercial premises and service centres. The rates applied for such licenses are based on the surface of the premises. AKKA/LAA enjoys a legal monopoly in Latvia.

9.

In 2008, the Konkurences padome (Competition Council, Latvia) imposed a fine on AKKA/LAA for abuse of a dominant position, since the latter had applied excessively high rates in respect of the remuneration of authors. The amount of the fine imposed was calculated by reference to the collecting society’s turnover, after deduction of that which had been transferred to authors as remuneration.

10.

Subsequently, in 2011, AKKA/LAA adopted new rates for authors’ remuneration, in relation to which the Competition Council commenced proceedings in 2012. To assess whether the rates were justified, that authority compared them both with the rates applicable in neighbouring Lithuania and Estonia, which were considered relatively similar to Latvia regarding consumption habits, economy and gross domestic product, and — by way of illustration — with the rates in other Member States, taking account of a purchasing power parity index based on the gross domestic product (‘the PPP index’). That authority ascertained that the applicant’s rates were appreciably higher than (in certain segments, even double) the rates in force in both neighbouring countries and among the highest in the Union, exceeding by between 50% and 100% the average level of rates in the Union. The Competition Council considered that those rates, in so far as they exceeded appreciably those laid down in the neighbouring countries, were not justified and, in addition, AKKA/LAA could not objectively justify them.

11.

Consequently, by decision of 2 April 2013 (‘the contested decision’), the Competition Council declared that AKKA/LAA’s conduct constituted a breach of the prohibition laid down in Article 13(4) of the Competition law and in point (a) of the second paragraph of Article 102 TFEU, and imposed a fine on it. For the calculation of the fine it took account of AKKA/LAA’s turnover, but this time including the sums collected in respect of authors’ remuneration which had been paid to the authors. The Competition Council indicated that the turnover of entities such as collecting societies must, for the purposes of competition law, be calculated in accordance with the same principles applied to companies with a share capital, so as to guarantee that determination of the fine does not differ depending on the legal form of the economic operator.

12.

By judgment of 9 February 2015, the Administratīvā apgabaltiesa (Regional Administrative Court, Latvia) partially upheld the claim: it declared that the conclusion that unjustifiably high rates had been applied was correct, but annulled the contested decision as regards the fine imposed and, on the basis of the principles of legality and equality, required the Competition Council to recalculate the fine to be imposed on AKKA/LAA, without including in the turnover the sums collected in respect of authors’ remuneration. Each party brought an appeal against that judgment before the Augstākā tiesa (Supreme Court, Latvia).

13.

Entertaining doubts as to the interpretation of Article 102 TFEU, the referring court decided to stay the proceedings and refer the following questions to the Court:

14.

Written observations have been submitted by AKKA/LAA, the German, Spanish, Latvian and Netherlands Governments and the Commission. AKKA/LAA, the Spanish and Latvian Governments as well as the Commission also presented oral argument at the hearing on 8 February 2017.

15.

This entire case centres on an alleged abuse consisting in the application of unfair prices under point (a) of the second paragraph of Article 102 TFEU. It therefore seems useful to recall briefly the Court’s case-law on that provision.

16.

In United Brands, ( 2 ) as well as in several subsequent decisions, ( 3 ) the Court regarded as contrary to what is now Article 102 TFEU the application of a price which was excessive because it had no reasonable relation to the economic value of the product supplied. Accordingly, only ‘disproportionate’ or ‘exorbitant’ prices could be in breach of that provision. ( 4 ) The Court laid down a two-step analysis for that purpose.

17.

The first step in the analysis is to determine whether there is an excess — that is a significant difference — between the price actually charged by the dominant undertaking in the relevant market and the price which that undertaking would hypothetically have charged had there been effective competition in the market (‘the benchmark price’). ( 5 )

18.

The Court has acknowledged that there may be different methods of determining whether the price is excessive. ( 6 ) For example, when possible and appropriate, a comparison can be made between the sale price and the cost of production. ( 7 ) This method appears based on the idea that there is a threshold price which guarantees a satisfactory margin ( 8 ) with respect to costs and that, beyond that threshold, the price charged by a dominant undertaking is excessive. ( 9 ) The focus of the analysis is thus on the margins (or on the profitability) made by the dominant undertaking in the sale of the products or services in question.

19.

In other cases, the Court has made a comparison between, on the one hand, the price charged for the product in question by the dominant undertaking and, on the other hand, the prices charged in the same market by non-dominant undertakings (comparison across competitors) ( 10 ) or by the same dominant undertaking at different points in time (comparison across time), ( 11 ) or the prices charged in other geographic markets by the same dominant undertaking ( 12 ) or by other undertakings (geographic comparison). ( 13 ) The underlying idea is that if the selected products or geographic markets are sufficiently homogenous, a comparison of the prices can be meaningful. ( 14 ) Likewise, the pricing patterns of an undertaking over time may also provide useful clues.

20.

Once it has been ascertained, by virtue of one or more of those methods, that a significant difference exists between the price actually charged by the dominant undertaking and the benchmark price, one must determine the extent to which that actual price is unfair, either in itself or when compared to competing products. ( 15 )

21.

That second step in the analysis is to investigate whether the difference in price is merely the result of an abusive use of market power by the dominant undertaking, or the consequence of other legitimate reasons.

22.

Only if no valid justification exists for the difference between the benchmark price and the actual price imposed by the dominant undertaking on its customers may a price be considered ‘unfair’ within the meaning of point (a) of the second paragraph of Article 102 TFEU.

23.

The Court has applied that two-step test in order to determine when a price is excessive, and thus unfair for the purposes of Article 102 TFEU, also in cases which concerned — as in the main proceedings — the conduct of collecting societies. In those cases, the Court has held that ‘when an undertaking holding a dominant position imposes scales of fees for its services which are appreciably higher than those charged in other Member States and where a comparison of the fee levels has been made on a consistent basis, that difference must be regarded as indicative of an abuse of a dominant position. In such a case it is for the undertaking in question to justify the difference by reference to objective dissimilarities between the situation in the Member State concerned and the situation prevailing in all the other Member States.’ ( 16 )

24.

It is against that background that I shall analyse the legal issues raised by the referring court.

25.

By its first question, the referring court asks the Court whether the conduct of a collecting society which is charged with the task of collecting remuneration also in respect of works of foreign authors may affect trade between Member States for the purposes of Article 102 TFEU.

26.

At the outset, I would recall that, according to settled case-law, the interpretation of the condition relating to effects on trade between Member States contained in Articles 101 and 102 TFEU must be based on the purpose of that condition, which is to define, in the context of the law governing competition, the boundary between the areas respectively covered by EU law and by the law of the Member States. Thus, EU law covers any agreement or any practice which is capable of constituting a threat to freedom of trade between Member States in a manner which might harm the attainment of the objectives of a single market between the Member States, in particular by sealing off domestic markets or by affecting the structure of competition within the single market. If an agreement, decision or practice is to be capable of affecting trade between Member States, it must be possible to foresee with a sufficient degree of probability, on the basis of a set of objective factors of law or of fact, that they may have an influence, direct or indirect, actual or potential, on the pattern of trade between Member States in such a way as to cause concern that they might hinder the attainment of a single market between Member States. ( 17 )

27.

That said, the mere fact that the conduct of an undertaking in a dominant position relates only to the marketing of products or services in a single Member State is not sufficient to preclude the possibility that trade between Member States might be affected. Indeed, such conduct may have the effect of reinforcing the partitioning of markets on a national basis, thereby holding up the economic interpenetration which the Treaty is designed to bring about. ( 18 )

28.

In the present case, as the referring court itself notes, the pricing policy implemented by AKKA/LAA concerns also works of foreign authors and, therefore, does affect the diffusion of those works in Latvia. Since that body enjoys a legal monopoly, its choices as regards whether, how and at what price it allows reproduction of protected works have an inevitable impact both on consumers’ patterns of conduct in Latvia and on copyright owners’ decisions with regard to that country’s market.

29.

Indeed, the Court has consistently found that EU competition rules apply to collecting societies’ activities consisting in licensing musical works, despite the fact that those activities were confined to only one Member State. ( 19 )

30.

The fact, alluded to in the order for reference, that in 2013 the General Court ( 20 ) partly annulled a Commission decision (‘the CISAC decision’), ( 21 ) adopted in the context of a proceeding under Article 101 TFEU, which had been addressed to 24 collecting societies (including AKKA/LAA), has no bearing in that regard. The General Court annulled the CISAC decision on the grounds that the Commission had not proved to the requisite legal standard the existence of a concerted practice between the collecting societies. Nothing in the General Court’s judgments concerns the issue whether the collecting societies’ conduct was capable of affecting trade between Member States.

31.

It must, therefore, be concluded that the conduct of a collecting society which is charged with the task of collecting remuneration also in respect of works of foreign authors may, despite being carried out in just one Member State, affect trade between Member States for the purposes of Article 102 TFEU.

32.

By its second question, the referring court essentially asks whether, in the situation in the main proceedings, it was appropriate and sufficient for the national competition authority to draw a comparison between the rates in the national market in question and the rates in neighbouring markets.

33.

This second question concerns — just like the third, fourth and fifth questions — the first step in the analysis alluded to in points 17 to 19 above: the assessment of whether an excess exists between the price actually charged by the dominant undertaking in the relevant market and the benchmark price. To recall, the latter is the price which that undertaking would hypothetically have charged had there been competition in the market.

34.

Clearly, this second question goes to the heart of the issues raised in the present proceedings since it requires the Court to clarify what methods and criteria fall to be applied by the competition authorities in order to determine the benchmark price. Before examining that aspect in detail, I would once again call to mind that the case in the main proceedings concerns alleged unfair pricing in a situation involving a legal monopoly.

35.

As I have explained in points 18 and 19 above, the Court has left the EU and national competition authorities a certain margin of manoeuvre with respect to the methodology that may be followed to determine an excessive price. For the following reasons, that is, to my mind, a very reasonable approach.

36.

It can be safely stated that, at the current stage of legal and economic thinking, there is no single method, test or set of criteria which is generally accepted in economic writings or across jurisdictions for that purpose. Different authorities as well as lawyers and economists have suggested a number of methods of analysis (as well as a variety of criteria, tests or ‘screens’) to that end. However, in point of fact, each of those methods reveals some inherent weaknesses.

37.

In the first place, none of those methods can be used in all circumstances, since their suitability (and, at times, the very possibility of applying them) depends very much on the specific features of each case. To give but one example, a cost-price comparison makes little sense with regard to the supply of certain intangible goods such as — as is the case in the main proceedings — copyrighted musical works.

38.

In the second place, the information required to perform the operations necessary to calculate the benchmark price may be missing or incomplete, or its value controversial. For instance, identifying costs and linking them to a particular product is highly complex in most types of businesses and for many undertakings. ( 22 ) Calculating profit margins is thus a rather uncertain exercise. It should not be overlooked that accounting standards and rates may change across industries or countries, due to different legal provisions or accounting conventions, and, furthermore, they may not always reflect the relevant economic concepts. ( 23 )

39.

In the third place, comparing prices across different geographic markets, competitors and/or time periods also presents risks. Markets are rarely so homogenous that a meaningful comparison can be made immediately and automatically. A number of ‘adjustments’ to the data which emerges from the market(s) used as a point of comparison may be necessary before that data can be used to determine the benchmark price.

40.

To start with, so far as concerns geographic comparisons, elements such as — to name but a few — domestic taxes, the particular characteristics of the national labour market and local consumers’ preferences may significantly affect the final prices of the relevant product or service. ( 24 ) With regard to comparisons across competitors, it should not be overlooked that differences in prices may simply reflect different qualities: a more expensive product may objectively be (or be merely perceived to be) of superior quality.

41.

Finally, as regards comparisons over time, one must be mindful of the fact that changes in factors which may affect the ultimate price of a product or service can occur rather quickly in the market. Those factors may concern legitimate business strategies (for example, an undertaking might decide to try to penetrate a new market and, for some time, charge a very low price, thus accepting minimal margins); an increase in costs (due to external factors such as changes in local taxation or borrowing costs, or to business decisions of the undertaking itself such as choices regarding advertising campaigns or research and development); or even consumer preferences (for example, shifts in the perception of a product by the customers in response to new marketing strategies). All those factors may lead to (usually legitimate) sudden and significant changes in prices.

42.

Because of those limitations, antitrust authorities and economists generally agree that the exercise consisting of determining the benchmark price in a case of possible excessive pricing carries a high risk of producing both type I errors (or false positives: a price is mistakenly considered to be above the competitive price) and type II errors (or false negatives: a price is mistakenly considered not to be above the competitive price). ( 25 )

43.

Thus, in the absence of an ubiquitous test and given the limitations inherent in all existing methods, it is in my view crucial that in order to avoid (or, more correctly, to minimise) the risk of errors, competition authorities should strive to examine a case by combining several methods among those which are accepted by standard economic thinking and which appear suitable and available in the specific situation. It seems to me that those which can be found in the Court’s case-law (and that have been illustrated in points 18 and 19 above) may serve that purpose. ( 26 )

44.

The choice to combine several methods is, in fact, the approach that a number of antitrust authorities have followed worldwide: for example, the UK Office of Fair Trading (OFT) has done so in the Napp case. ( 27 ) It is also consistent with suggestions made in international discussion fora of those authorities ( 28 ) as well as in contemporary economic literature. ( 29 )

45.

It is true that such an approach has been criticised on the ground that the combined application of several imprecise methodologies, even where producing mutually consistent results, may not lead to a more reliable conclusion. ( 30 ) Admittedly, the weaknesses of one method are not necessarily remedied by applying another equally weak method. Yet, if the methods are applied independently of each other, a given limitation inherent to one of them would not affect the results obtained through the use of other methods. Accordingly, provided that the methodologies used are, in themselves, not flawed, and that they are all applied with rigour and objectivity, the convergence of results may be taken as an indicator of the possible benchmark price in a given case.

46.

That said, there may be cases in which only one of those methods of determining the benchmark price may be available or suitable. In those cases, I find it of the utmost importance that the authority consider other indicators which may corroborate or, conversely, cast doubt on the result of that method.

47.

The following seem to me to be relevant indicators.

48.

First, a price cannot easily be set significantly above the competitive level where the market is not protected by high barriers to entry or expansion. Otherwise, as mentioned above, the market should, in principle, be able to self-correct in the short to medium term: high prices should normally attract new entrants or encourage existing competitors to expand. That is why — as stated in the beginning of this Opinion — I am convinced that unfair prices under Article 102 TFEU can only exist in regulated markets, where the public authorities exert some form of control over the forces of supply and, consequently, the scope for free and open competition is reduced. Obviously, the higher and longer-lasting the barriers created by the legislature, the more a dominant undertaking should be able to exercise its market power.

49.

Second, a price significantly in excess of a competitive price is more unlikely to occur in markets where there is a sectoral regulator whose task is, inter alia, to fix or control prices charged by the undertakings active in that sector. Sectoral authorities are clearly better-equipped than competition authorities to oversee prices and, where necessary, act to remedy possible abuses. ( 31 ) It would seem, therefore, that antitrust infringements in those situations should be mainly confined to cases of error or, more generally, to regulatory failures: cases where the sectoral authority should have intervened and erroneously failed to do so.

50.

Third, an undertaking with market power is evidently less able to leverage its position when negotiating with powerful buyers. To give an example, so far as concerns licences for the use of copyrighted musical works, the negotiating position of small shops is likely to be different from that of international platforms (such as Spotify) or groups of large and sophisticated undertakings (such as Hollywood majors). The size and financial strength of an undertaking (or group of undertakings) might indeed have a significant weight in the negotiations. However, the extent to which the licenced products constitute an important (or even indispensable) input for the customers’ business may also be of great importance in that context.

51.

Clearly, there may be other factors which may be relevant, depending on the specific circumstances of each case.

52.

Concluding on this point, it seems important to make the following two observations. First, I would recall that it is for an antitrust authority to prove an infringement of EU competition rules. ( 32 ) Second, it is settled case-law that a principle such as the presumption of innocence applies to undertakings investigated for possible infringements of EU competition law. ( 33 )

53.

As a result, in my view, a lack of reliable data or the complexity of the operations involved in the calculation of the benchmark price (or in corroborating it) cannot justify an incomplete, superficial or dubious analysis by a competition authority. In other words, difficulties encountered by an authority when carrying out an assessment cannot be to the detriment of the undertaking being investigated.

54.

Regardless of the specific situation in a given case, the method(s) applied and the other indicator(s) examined must give the authority a sufficiently complete and reliable set of elements which point in one and the same direction: the existence of a difference ( 34 ) between the (hypothetical) benchmark price and the (actual) price charged by the dominant undertaking in question.

55.

It is against that background that I shall examine the specific aspects of the case at issue in the main proceedings.

56.

In the contested decision, the Competition Council decided to compare the rates applied by AKKA/LAA with those applied by similar bodies active in other geographic markets. The referring court asks whether that method was, in the case at hand, appropriate and sufficient.

57.

As mentioned in points 19 and 23 above, the geographic comparison method has — in principle — been recognised as valid by the Court. It was, furthermore, endorsed in cases which concerned precisely the conduct of collecting societies.

58.

I thus agree with the German, Spanish, Latvian and Netherlands Governments and the Commission that a geographic comparison between the prices practiced for the very same service by different bodies in different Member States may, in a situation such as that in the main proceedings, ( 35 ) be an appropriate method of determining the benchmark price for the purposes of Article 102 TFEU.

59.

That is so, obviously, only if the authority has applied the method correctly.

60.

Whether a given method has been applied correctly in a specific case is clearly something which is, as a matter of principle, for the competent national courts to determine. However, the Court may, where possible, give guidance to those courts so that they can interpret and apply Article 102 TFEU properly and consistently.

61.

In that regard, I am of the view that an authority should, first, select the Member States of reference according to objective, appropriate and verifiable criteria.

62.

According to the referring court, the Competition Council chose the neighbouring countries of Lithuania and Estonia because they are relatively similar to Latvia regarding consumption habits, economy and citizens’ welfare (gross domestic product) and they also share the same historical and cultural heritage.

63.

Contrary to the view expressed by AKKA/LAA, I find those criteria to be objective and verifiable. They appear, in addition, relevant insofar as they are meant to ensure that the markets are homogenous on both the demand and the supply side. It is indeed crucial, in this context, to take into account the following two factors which, in my opinion, could affect the economic value of the service provided by AKKA/LAA: (i) the capacity and willingness of AKKA/LAA’s customers to pay for the service received; and (ii) the economic benefit that AKKA/LAA’s customers may derive from that service when, in turn, they supply products or services to their own customers.

64.

It is however for the referring court to verify that the alleged similarities between Latvia on the one hand, and Lithuania and Estonia on the other, are real and truly relevant for the analysis made by the Competition Council.

65.

Importantly, that court is also to ensure that no other Member State, despite not being a neighbouring country, ( 36 ) fulfils the criteria adopted by the Competition Council. In other words, as argued by the German Government, the Augstākā tiesa (Supreme Court) is also to verify that the Competition Council has not excluded countries arbitrarily or, worse still, because they produced data which did not ‘suit’ its case.

66.

In that regard, the referring court indicates that, in its decision, the Competition Council has also considered, by way of illustration, the rates applied in other Member States (both individually and to calculate the EU average), taking account of the PPP index based on gross domestic product. The results of that analysis appear to confirm the conclusions reached by examining the markets of Lithuania and Estonia.

67.

That ‘enlargement’ of the group of countries with which the Latvian market has been compared is of the utmost importance. A comparison limited to only two countries — as homogenous as they may be with Latvia — might not provide reliable results. Indeed, as AKKA/LAA points out, any atypical factor which may exist in either of those two markets would have a particularly significant effect on the calculations made by the competition authority. To my mind, the sample of countries for a comparison must be as broad as possible. ( 37 )

68.

That said, any meaningful difference between the relevant Member State and the other Member States chosen for a comparison should be accounted for. As mentioned above, the Court has indeed made clear that a comparison across countries is possible if done on a consistent basis. The referring court should thus check that the necessary adjustments have been made, so as to take into account the differences existing between the different countries.

69.

The final issue to examine in order to provide an answer to the referring court is whether the geographic comparison method followed by the Competition Council was sufficient for the purposes of establishing the benchmark price.

70.

Again, that is an issue which is in principle for the national court to decide. In order to provide guidance to the referring court, however, I would note the following.

71.

The referring court should first verify whether alternative methods of determining the benchmark price could also have been employed alongside the geographic comparison. Subject to verification by the national court, my impression is that, probably, certain other methods may not have been available or suitable.

72.

First, a cost-price analysis seems impossible to carry out in the situation which was the subject matter of the contested decision (what is the cost of composing a musical work?). ( 38 )

73.

In this context, however, I observe that the Netherlands Government suggests a different type of analysis which focuses instead on the remuneration that, by virtue of the tariffs applied by AKKA/LAA, the authors of the licensed works actually receive. In fact, there are provisions in a number of EU directives which concern precisely the remuneration that copyright holders are to receive for the exploitation of their works.

74.

Article 8(2) of Directive 92/100/EEC, ( 39 ) for example, states that users must pay to copyright holders an ‘equitable remuneration’ if a phonogram published for commercial purposes, or a reproduction of such a phonogram, is used for broadcasting by wireless means or for any communication to the public. The Court has interpreted the concept of ‘equitable remuneration’ as one which enables ‘a proper balance to be achieved between the interests of performing artists and producers in obtaining remuneration for the broadcast of a particular phonogram, and the interests of third parties in being able to broadcast the phonogram on terms that are reasonable’. The Court has also indicated that whether the remuneration is equitable ‘is to be assessed, in particular, in the light of the value of that use in trade’. ( 40 )

75.

In addition, Article 16(2) of Directive 2014/26/EU, ( 41 ) which concerns licensing by collective management organisations, provides that ‘rightholders shall receive appropriate remuneration for the use of their rights’. ( 42 )Although Directive 2014/26 is not applicable in the main proceedings ratione temporis, it may nonetheless be relevant since the concept of ‘appropriate remuneration’ seems akin to that of ‘equitable remuneration’ appearing in Directive 92/100. In the view of the Netherlands Government, tariffs which lead to a remuneration which is equitable or appropriate cannot be considered abusive under Article 102 TFEU.

76.

The approach of the Netherlands Government does seem appealing: if unfair prices are those which excessively exploit customers to the benefit of the dominant undertakings, it may be logical to consider that tariffs which do not strike a fair balance between the interests of the copyright holders and those of the customers may fall foul of Article 102 TFEU. The approach taken by the Court in Kanal 5 seems, indeed, to lend some support to that position. ( 43 )

77.

That said, I have doubts as to whether the legal frameworks established by Directives 92/100 and 2014/26 on the one hand, and Article 102 TFEU on the other, fully coincide: they pursue different objectives and respond to a different logic. The directives seek, inter alia, to ensure that authors and performers receive an adequate income as a basis for further creative and artistic work. ( 44 ) Conversely, Article 102 TFEU seeks to ensure that undertakings with a dominant position (including collecting societies) do not abuse their market power.

78.

At any rate, I am not sure that concepts such as ‘equitable’ or ‘appropriate’ remuneration could be of great assistance to a competition authority. They seem to me as vague as the concepts of ‘excessive’ or ‘unfair’ prices.

79.

Second, since AKKA/LAA is a legal monopolist, there are no similar services offered by competing undertakings in Latvia that could be used for a comparison. In addition, AKKA/LAA is not active outside Latvia. As concerns comparisons of tariffs practiced by AKKA/LAA across different periods, it is not clear whether that could have provided useful points of reference since the Competition Council had already found past tariffs excessively high.

80.

That said, it is for the referring court to check if other methods of determining the benchmark price that could have theoretically been used in combination with the comparison across different Member States were available and suitable. It is also for the referring court to verify whether the results at which the Competition Council arrived as regards the benchmark rates have been corroborated by additional indicators.

81.

On the basis of the above, I propose that the Court answer the second question as follows: in a situation such as the one in the main proceedings, it is, in principle, appropriate to draw a comparison between the rates in the market in question and the rates in other markets. It is, however, for the national court to verify, in the light of all the relevant circumstances, whether that comparison was, on the one hand, correctly carried out and, on the other hand, sufficient.

82.

By its third question, the referring court asks whether it is appropriate and sufficient to use the PPP index when comparing the rates charged by different collecting societies.

83.

The referring court explains that, when comparing the rates applied by AKKA/LAA in Latvia with those applied in 19 other Member States (that is to say the Member States other than the neighbouring countries), the Competition Council made use of the PPP index to ‘correct’ those rates.

84.

From the outset, I would call to mind once again that, in Tournier and Lucazeau, the Court stated that a geographic comparison of the rates may be possible provided that it is made ‘on a consistent basis’. ( 45 ) To my mind, the consistency of a comparison requires not only that the products and services in question must be the same or very similar, but also that the economic context in which those products and services are supplied must be broadly similar.

85.

However, it cannot be disputed that, within the Union, significant differences in price levels exist, meaning that for the same goods or service citizens pay different prices in various countries. Even when countries use the same currency, the purchasing power of consumers may vary.

86.

This is why, in line with the arguments put forward by the German, Spanish, Latvian and Netherlands Governments, I too am of the view that the PPP index can be a useful instrument to ensure that a comparison of the rates applied for the very same service in different countries is made on a homogenous basis.

87.

A PPP index is, in fact, commonly used in economic studies — including by bodies such as Eurostat, the OECD or the World Bank — when a cross-country comparison of, for example, living standards must be made. To that end, PPP currency conversion rates are applied in order to convert economic indicators from national currency into an artificial common currency, called the Purchasing Power Standard (PPS), which equalises the purchasing power of different national currencies and enables meaningful comparisons between countries to be made. Those operations thus permit the adjustment of data that is to be compared, according to the different price levels existing in the various countries.

88.

AKKA/LAA and the Commission, however, object that such an instrument could be useful only for that part of the rates which is kept by the collecting society, and not for the part of the rates which constitutes the remuneration of the copyright owners.

89.

I disagree.

90.

It is not only the costs of the collecting society that are affected by the economic situation of the country in which it operates. The capacity and, to a certain extent, willingness to pay displayed by the collecting society’s customers (in the present case, the shops) and, in turn, by the latter’s customers (in the present case, the customers of the shops) are also influenced by the citizens’ standard of living and purchasing power. To put it very simply: if EUR 1 in a country does not equal EUR 1 in another country, that is true irrespective of whether that money goes to finance the costs of the collecting society or to remunerate authors. After all, if the key aim of the analysis is to identify the economic value of a given transaction, that assessment cannot be made in abstracto, but must necessarily take into account the economic and financial context in which the transaction takes place.

91.

The use of a PPP index may thus be a useful instrument to analyse the rates applied by a collecting society in their entirety. There is no need to distinguish between the different components of those rates.

92.

Accordingly, I conclude that, to the extent that an authority undertakes a geographic comparison of the rates applied by various collecting societies, the different economic situation of the countries in which those collecting societies operate must be taken into account. The use of a PPP index seems to me an appropriate instrument to that end.

93.

Whether that instrument is sufficient depends, however, on whether the other factors which may affect the final price of a product or service in a given country are also taken into account. Indeed, there may be other factors — including non-macroeconomic ones — which may affect the structure of demand in a country. In particular, in a case such as that in the main proceedings, it seems to me that whether and to what extent customers of the collecting societies (for example, shops) in a given country are able to increase their business activities by reason of the public reproduction of music in their premises is key in that regard.

94.

Those are, at any rate, factors that may also be examined in the second step of the legal analysis laid down in the Court’s case-law. I will thus address that issue when examining the sixth question referred.

95.

In the light of the above, the answer to the third question referred should be that the use of a PPP index may be appropriate when comparing the rates charged by different collecting societies. Whether that instrument is sufficient depends on whether the other factors which may affect the final price of a product or service in a given country are also taken into account.

96.

By its fourth question, the referring court asks whether, in a situation such as that in the main proceedings, the comparison of the rates charged by different collecting societies should be made for each separate segment of the market, or in relation to the average level of the rates.

97.

The answer to this question is, to my mind, rather straightforward.

98.

Whether a given conduct of one or more undertakings breaches Article 101 or 102 TFEU is to be determined by looking at the relevant market.

99.

Thus, assuming that each separate segment of the market (meaning a category of users determined on the basis of the surface exploited commercially) is a relevant product market for the purposes of Article 102 TFEU — which is for the referring court to verify — a comparison of the rates charged by different collecting societies should be made for each separate segment of the market.

100.

By its fifth question, the referring court seeks guidance on the circumstances in which a difference in price can be considered excessive pursuant to point (a) of the second paragraph of Article 102 TFEU.

101.

At the outset, let us start by recalling the economic rationale of the unfair pricing abuse: when a dominant undertaking applies prices above competitive levels, there is an inefficient allocation of resources and consumer welfare is reduced (part of the welfare is transferred to the dominant company, whereas part is simply lost). Accordingly, from a theoretical point of view, any deviation from the competitive price in a regulated market might justify an intervention of the competition authorities. Indeed, any difference between the benchmark price and the actual price implies a certain loss in consumer welfare that would not have been there had the market been competitive.

102.

However, such an approach would, for a competition authority, neither be realistic nor advisable.

103.

First, as explained in points 36 to 42 above, the calculation of a benchmark price is a rather complex and uncertain exercise. Were a competition authority to intervene in respect of any difference — however small — between those two prices, the risk of having false positives would simply be too high. That is not only a problem because a large fine may be imposed on the undertaking responsible, but also because neutral — or possibly pro-competitive — conduct might be prohibited. In that regard, it has been correctly argued that type I errors in competition decisions concerning unilateral conduct involve a much larger cost for the society than type II errors: ‘the economic system corrects monopolies more readily than it corrects judicial errors … A practice once condemned is likely to stay condemned, no matter its benefits. A monopolistic practice wrongly excused will eventually yield to competition though, as the monopolist’s higher prices attract rivalry.’ ( 46 )

104.

Second, because of those difficulties and uncertainties, it must also be acknowledged that it may often be difficult for a dominant undertaking to estimate in advance, with a sufficient degree of likelihood, where the line between a legitimate competitive price and a prohibited excessive price may be drawn. Thus, for reasons of legal certainty, that threshold cannot be set too close to the benchmark price.

105.

Third, a strict approach would require competition authorities essentially to become price regulators which ought continuously to monitor and intervene in (potentially all) regulated markets. Clearly, unlike sectoral authorities, competition authorities have neither the resources nor the expertise to do that. ( 47 ) Moreover, the loss of consumer welfare may at times be minor and not justify a complex, time-consuming and costly intervention by the public authorities. Indeed, how consumers react to a price increase differs widely from market to market, and not even a monopolist can set prices independently of its customers. ( 48 ) Thus, the degree of damage made to consumers’ welfare by high prices may vary.

106.

That is why — in line with the approach adopted by the relevant authorities and courts both at the EU level and at the Member States level, and as suggested in economic writings — I take the view that a price can be qualified as excessive under Article 102 TFEU only if two conditions are fulfilled: it ought to be both significantly and persistently above the benchmark price.

107.

As regards the first aspect, I would emphasise that not any price difference should be regarded as relevant under Article 102 TFEU but only important deviations. That approach has been expressly endorsed by the Court: for example, in Tournier and Lucazeau, the Court referred to scales of fees ‘appreciably higher’ than those to which they are compared. That viewpoint is also widely supported in economic writings. ( 49 )

108.

As concerns the second aspect, the fact that the price of a given product or service is sporadically above the benchmark price is, to my mind, of little relevance. The existence of periods of high prices alongside periods of low prices is, in economic writings, considered ‘consistent with a well-oiled competitive market’. ( 50 ) Thus, a price which varies continuously and is only from time to time above competitive levels is, in my view, unlikely to raise serious competition concerns. Only when a price remains (or is recurrently) above the benchmark price for a substantial period of time may that price be abusive under Article 102 TFEU. Support for that approach can be found in General Motors. ( 51 )

109.

All the above raises the question: how significant and how persistent must that difference be to justify an intervention under Article 102 TFEU?

110.

That is by no means an easy question to answer. Existing case-law of the Court does not give very precise guidance on this issue. Nor can clear patterns be found in the practice of national authorities or in economic literature. ( 52 )

111.

That is not at all surprising. Indeed, as the German Government and the Commission argue, it is impossible to set, a priori and in abstracto, precise thresholds which could be applicable to all circumstances. A given difference in price could be more or less significant, under Article 102 TFEU, depending on the product or service in question and the characteristics of the market.

112.

On that point, I would only add the following two considerations. On the one hand, an authority should intervene under Article 102 TFEU only when it feels sure that, regardless of the limitations and uncertainties surrounding the calculation of the benchmark price, the difference between that price and the actual price is of such a magnitude that almost no doubt remains as to the latter’s abusive nature. On the other hand, the more significant the difference between the benchmark price and the actual price, and the longer the period in which that high price is applied, the easier it should be for an authority to discharge its burden of proof. ( 53 )

113.

The answer to the fifth question should thus be as follows: only prices which are significantly and persistently above the benchmark price may be considered to fall foul of Article 102 TFEU.

114.

By its sixth question, the referring court asks how a collecting society may prove the fair nature of the rates charged.

115.

In essence, this question invites the Court to give further explanations on the second step of the legal analysis required under point (a) of the second paragraph of Article 102 TFEU.

116.

As mentioned above, the fact that there is an excess — even a substantial one — between the benchmark price and the actual price is not sufficient for that price to be considered automatically unfair under Article 102 TFEU, or in any event to justify intervention under that provision.

117.

High prices are normally not per se abusive. On the contrary, they fulfil an important function in the competitive process. As the US Supreme Court stated in Trinko: ‘the mere … charging of monopoly prices, is not only not unlawful; it is an important element of the free-market system. The opportunity to charge monopoly prices — at least for a short period — is what attracts “business acumen” in the first place; it induces risk taking that produces innovation and economic growth. To safeguard the incentive to innovate, the possession of monopoly power will not be found unlawful unless it is accompanied by an element of anticompetitive conduct’. ( 54 )

118.

The focus of this second step in the analysis must thus be on the dominant undertaking’s conduct and its economic motives. In particular, the objective reasons behind its pricing policy are highly relevant.

119.

The Court has, in United Brands as well as in subsequent case-law specified that a price may be unfair ‘in itself or when compared to competing products’. ( 55 )

120.

What are the reasons behind those alternative conditions? ( 56 )

121.

The first of those two conditions (a price unfair in itself) is meant to cover those instances in which the unfairness of a price can be determined without the need to make any comparison with similar or competing products. The particularly high price in itself reveals the abuse.

122.

That may be the case, for example, of prices which are charged to customers which, however, do not receive any product or service in return. For example, in Merci Convenzionali Porto di Genova, the Court considered incompatible with (now) Article 102 TFEU national legislation which induced an undertaking to which special rights were granted to, inter alia, demand payment for services which had not been requested. ( 57 ) Similarly, in Grüne Punkt, the Court upheld a Commission decision which had found a dominant undertaking’s power to claim payments from its contractual partners for services which it had not provided to be in breach of (now) Article 102 TFEU. ( 58 )

123.

It may also be the case of situations in which a dominant undertaking sets a price particularly high because, in reality, it is not interested in selling the product or service in question but intends to pursue a different, anticompetitive, aim. That situation may be observed in the General Motors and British Leyland cases. ( 59 ) In those cases the undertakings in a dominant position (car manufacturers) had set very high prices for conducting technical inspections and issuing certificates of conformity. The reason was — as the Court explained in its judgments — that the car manufacturers wished to curb parallel imports into the United Kingdom by neutralising the more favourable level of prices applying in other areas of the then Community. There was clearly no reasonable relation between the prices charged by the car manufacturers and the quantity and quality of the services provided to importers.

124.

The second of those conditions (price unfair when compared to competing products) is, often, a ‘sanity-check’ of the assessment made with regard to the benchmark price: there may be relevant factors which were either overlooked in that context, or were consciously not taken into account because they were not easily quantifiable in financial terms.

125.

Indeed, there may be a variety of reasons — possibly legitimate — why an undertaking may price a given product or service above the price which was calculated by the authority as being the (hypothetical) competitive price. That means that, even if the market were to be competitive, the price applied by the dominant undertaking would perhaps still not correspond to the benchmark price, its products or services having a higher economic value.

126.

Those reasons for a higher price may concern, in particular, the production and marketing of the product or service in question, but they may also pertain to customers’ demand for that product or service.

127.

As regards the first aspect, I would emphasise that the costs for a dominant undertaking to produce and market its product or service in question may be higher than those incurred by the other undertakings, which are not dominant or operate in other product or geographic markets. An authority ought to consider not only direct and indirect production costs for the product or service in question and the cost of capital, but also all types of overheads (including, for example, advertisement, research and development, and so forth). ( 60 ) Even if a dominant undertaking cannot simply justify its higher prices by reason of a possibly inefficient or uneconomical cost structure, ( 61 ) the actual costs incurred by that undertaking are obviously of crucial importance in that regard. Certain types of costs which a given undertaking may have incurred may not be immediately evident or easily imputable to the supply of a given product or service (for example, abortive research and development), ( 62 ) but nonetheless may not be discounted. A different approach would seriously risk discouraging investment and innovation.

128.

As concerns the second aspect, I would point out that the economic value of the goods or service supplied by a dominant undertaking may, in the eyes of the customers, be higher than the benchmark price. Again, there may be a variety of reasons for that: for instance, the goods or service in question may be (or be merely perceived to be, perhaps for reasons relating to advertising or branding investment costs) of superior quality. Some features of the product or service may be regarded as particularly valuable by customers (or certain groups of customers), in spite of the fact that they are not reflected on the cost side. In those cases, the additional benefits or advantages provided to customers justify a higher mark-up over costs. ( 63 ) In this context, I note that the Commission’s practice seems to follow that approach. ( 64 )

129.

Having said that, I would add that, so far as concerns the situation in the main proceedings, the key issue seems to be the following: the demand for licences by AKKA/LAA’s customers such as shops or other similar undertakings is a direct function of the economic benefits they can draw from those licenses. Accordingly, higher rates in Latvia could be justified if it were to be proven that the benefits that AKKA/LAA’s customers derive from music reproduction were to be larger than those derived by the same type of customers in other countries. For example, it cannot be ruled out that, because of different purchasing habits and cultural traditions, shops and other commercial activities in some countries may increase their business more than in other countries by reason of the public performance of music in their premises. In those circumstances, the economic value of the licences granted by the collecting society would naturally be higher in the former countries than in the latter ones.

130.

I recognise that such an aspect may not be easy to investigate. That is the reason why other indicators (such as the citizens’ purchasing power, and the countries’ gross domestic product) are often used to determine whether and to what extent two or more countries are comparable from the point of view of the economic situation.

131.

In conclusion, it is only when no rational economic explanation —– other than the mere capacity and willingness to use market power even when abusive — can be found for the high price applied by a dominant undertaking that that price may be qualified as abusive under Article 102 TFEU.

132.

Before concluding on this matter, there is a last point which deserves some attention. The two-step analysis laid down in the Court’s case-law as regards infringements of point (a) of the second paragraph of Article 102 TFEU also has a procedural aspect.

133.

Indeed, as mentioned in point 23 above, the Court has repeatedly stated that, once a price has been found to be in excess of the benchmark price, it is ‘for the undertaking in question to justify the difference by reference to objective dissimilarities’ between the products or services compared. ( 65 )

134.

That statement of the Court should, in my view, be read against the background of a well-established line of case-law according to which, while it is for the authority to discharge its burden of proving that all conditions required for a finding of an infringement of Article 102 TFEU are fulfilled, ( 66 ) dominant undertakings are to have the possibility of demonstrating an objective justification for their conduct. ( 67 )

135.

Thus, once an authority has recorded an excess between the actual price and the benchmark price, it is for the dominant undertaking in question to provide the authority with possible justifications for the (real or apparent) higher price.

136.

That is reasonable: the investigating authority often lacks information which may be necessary to assess whether a price that appears to be above the competitive price does not, in reality, merely reflect the higher value of the underlying transaction. Such information may concern, inter alia, the dominant undertaking’s cost structure, its pricing policies, the structure of demand in the relevant market, and so forth.

137.

The authority has to carefully and impartially review the factors put forward by the undertaking in question, before it decides on the possible unfairness of the price.

138.

In the present case, it means that it was, first, for the Competition Council to demonstrate to the requisite legal standard that the rates applied by AKKA/LAA were significantly higher than the competitive price. To that end, that authority was required to take into account, during an objective and thorough investigation, all the relevant facts in order to determine the correct benchmark price.

139.

It was then for AKKA/LAA to show the fair nature of the rates applied, in spite of the fact that they were higher than the benchmark price found by the Competition Council. AKKA/LAA could, for example, point to relevant factors mistakenly overlooked by the Competition Council when calculating the benchmark price; or in any event demonstrate that the economic value of the service supplied to its customers was higher than that provided by similar bodies in other Member States.

140.

In conclusion, I propose to answer the sixth question to the effect that a dominant undertaking may prove the fair nature of the prices applied on grounds, in particular, of higher production and marketing costs, or, more generally, of higher economic value of the product or service supplied.

141.

By its seventh question, the referring court asks whether, for the purposes of determining the fine to be imposed on a collecting society for an infringement of EU competition rules, the remuneration paid to authors should be excluded from the turnover of that body.

142.

On this point I find myself in agreement with the Spanish Government and the Commission: I do not see any reason why the remuneration paid to authors should be excluded from the turnover that is taken as a basis for the calculation of the fine imposed on a collecting society.

143.

In a number of judgments, and again recently in OSA, ( 68 ) the Court has stated that collecting societies should be regarded as undertakings for the purposes of EU competition rules. For those undertakings, the ‘total turnover’ — which is referred to, inter alia, in Article 23(2) of Regulation No 1/2003 and the Commission Guidelines ( 69 ) — includes the part of the rates which corresponds to the remuneration of the authors. It is irrelevant that that part is then paid to authors. In that sense, the remuneration paid to authors could be regarded as a ‘cost’ item for the collecting society.

144.

After all, if the relevant turnover were considered to be limited to only the part of the revenues that the collecting society is able to keep, the fine would be of a relatively low amount. That may raise doubts as to whether that fine would be sufficiently deterrent, proportionate in the light of the harm done to the consumers, and fair when compared to the fines imposed on other undertakings which have committed similar breaches of EU competition rules.

145.

In conclusion, I propose that the Court answer the questions referred for a preliminary ruling by the Augstākā tiesa (Supreme Court, Latvia) as follows: