–

first, errors in the classification of the conduct in question as a ‘single and continuous infringement’ of Article 101(1) TFEU and the extent of their participation in that infringement;

–

secondly, errors in the classification of that infringement as a ‘restriction by object’; and

–

thirdly, errors in determining the amount of the fines imposed.

Before addressing those three sets of pleas in law common to both actions, the Court first examines Crédit agricole’s plea in law alleging breach of the principle of the presumption of innocence.

Respect for the presumption of innocence

With regard to respect for the presumption of innocence, Crédit agricole argued that the Commission had wrongly assumed that the traders involved, and in particular its own trader, were aware of all the information exchanged on the permanent chat rooms to which they were connected, regardless of their active participation in those exchanges.

That complaint is rejected by the Court, which emphasises that the forums in question were characterised by the real-time delivery of messages to all those connected. In view of that particularity, the Commission was entitled to consider that Crédit agricole had been aware of the discussions held on those forums as soon as its trader had logged on, even if the trader had not actively participated in those discussions or even if he had had access to numerous other concomitant sources of information. That could only have been the case if Crédit agricole had demonstrated, by means of certain and precisely time-stamped evidence, that its trader had not in fact become aware of the offending message(s). Crédit agricole had not provided such proof. In that respect, the terms of the discussions in question differ from those which gave rise to the judgment in Eturas and Others. ( 2 )

By contrast, the Court finds that the Commission breached the principle of the presumption of innocence by fixing the starting point of Crédit agricole’s participation in the breach at the date of its trader’s first connection to the chat room in question using that bank’s identifiers, which occurred on 10 January 2013.

In order to accept that first connection as evidence of anticompetitive conduct marking the beginning of Crédit agricole’s participation in the breach, it was for the Commission to show that, on the very day of that first connection, the Crédit agricole trader had at least passively participated in an anticompetitive discussion. In the present case, neither the contested decision nor the documents before the Court show that messages of an anticompetitive nature were exchanged on the chat room at issue on 10 January 2013 after the Crédit agricole trader had logged on for the first time.

The participation of the applicants in a single and continuous infringement

As regards the pleas challenging the classification of the conduct at issue as a ‘single and continuous infringement’ attributable to the applicants, the Court observes, as a first step, that only conduct forming part of an ‘overall plan’ pursuing a single anticompetitive objective may be classified as a single and continuous infringement.

As regards the single nature of the infringement, the Court considers that the Commission correctly considered that the sole anticompetitive objective pursued by the traders of the banks concerned was to maximise the latter’s income while limiting the losses which could result from the uncertainty linked to the conduct of other traders.

Since the Commission has demonstrated to the requisite legal standard that the conduct adopted by the traders of the banks concerned between January 2010 and February 2013 formed part of an overall plan pursuing that single anticompetitive objective, the Court also considers that the prohibition imposed by Deutsche Bank in February 2013 on its traders to use permanent multilateral chat rooms did not prevent the traders of the banks concerned from achieving that objective. On that point, the Court states that the unique nature of an infringement results from the uniqueness of the objective pursued by the participants in the cartel. It was not disputed that the traders of the banks concerned had circumvented the prohibition addressed to the Deutsche Bank traders in February 2013 by means of a network of bilateral discussions, which functioned in the same way as the permanent multilateral chat rooms.

As regards the continuous nature of the infringement, the Court confirms that the context in which the cartel found supports the Commission’s conclusion that the banks concerned participated in a continuous infringement between January 2010 and March 2015. Although the exchanges between the traders of those banks became less frequent after February 2013, the fact remains that they continued their discussions of an anticompetitive nature on a recurring basis, freely exchanging information on their ongoing trading activities.

Crédit agricole’s argument that it did not participate in the infringement during certain periods is not such as to call into question the continuing nature of the infringement as a whole, since the interruptions relied on by Crédit agricole do not take account of the conduct of the other participants.

As regards imputability to the applicants of the single and continuous infringement, the Court points out, secondly, that that imputability must be assessed in the light of two factors, namely, first, their intentional contribution to the common objectives pursued by all the banks concerned and, secondly, their knowledge of the infringing conduct envisaged or implemented by those banks in pursuit of the same objectives or the fact that they could reasonably have foreseen it and had been prepared to accept the risk.

Having made that clarification, the Court rejects all the arguments put forward by the applicants in order to contest both their intentional contribution to the overall plan identified by the Commission and their knowledge of all the infringing conduct in question or, as the case may be, their ability to foresee it.

In that context, the Court notes that the Commission’s conclusion that Crédit agricole could, at the very least, reasonably have foreseen all of the infringing conduct of the other banks is corroborated in particular by the fact that, before taking up his duties at Crédit agricole, its trader had, as a trader at another bank, participated directly in the infringing conduct at issue.

On that point, the Court emphasises that knowledge acquired by an employee prior to his or her arrival in the service of a new undertaking and which he or she in fact makes available to that new employer may be regarded as knowledge shared by his or her new employer. Moreover, it is settled case-law that the Commission may rely on contacts prior to or subsequent to the period of the infringement in order to build up an overall picture and to show the preparatory stages of the cartel as well as to corroborate the interpretation of certain items of evidence.

In the light of the foregoing, the Court rejects all of the applicants’ complaints challenging, first, the classification of the conduct at issue as a ‘single and continuous infringement’ and, secondly, the imputability of that infringement to the applicants.

Classification of the conduct at issue as a ‘restriction by object’

Referring to the case-law of the Court of Justice, the General Court points out that, for the purposes of classifying the conduct at issue as a ‘restriction by object’, it was for the Commission to show that that conduct did not present an extremely high threshold of harm to competition, as Crédit agricole argued, but only a sufficient degree of harm to competition.

The Court also states that the assessment of the degree to which conduct is harmful to competition must be made in the light of the objective characteristics of that conduct and without regard to the particular situation of each undertaking that participated in it. Thus, the minor role played by one undertaking in a cartel is not such as to influence the classification of that cartel as a ‘restriction by object’ in respect of all of the undertakings that participated in it. For the same reasons, Crédit agricole cannot usefully rely on the fact that it did not take part in certain discussions to contest the classification of the conduct in question as a ‘restriction by object’.

In the light of those clarifications, the Court then rejects the applicants’ complaints alleging errors made by the Commission, first, in assessing the economic context of the conduct at issue, secondly, in assessing whether it was harmful to competition and, thirdly, in assessing whether it was justified by reason of its pro-competitive effects.

As regards, in the first place, the assessment of the economic context of the conduct at issue, the Court finds that, while in a complex market such as the present one the Commission cannot limit its analysis of that context to what is strictly necessary in order to conclude that there is a restriction of competition by object, the applicants have failed to demonstrate any inadequacy in the Commission’s analysis of the economic and legal context.

As regards, in the second place, the assessment of whether the conduct at issue was harmful to competition, the Court endorses the Commission’s conclusion that, on the secondary market in SSA bonds, the exchanges of commercially sensitive information between the banks concerned, all of which were ‘market makers’, ( 3 ) were sufficiently harmful to competition to contribute to the classification of the conduct examined, as a whole, as a ‘restriction by object’.

That conclusion cannot be called into question by Crédit agricole’s allegation that the secondary market in SSA bonds is a market in which there is a significant asymmetry of information between market makers, so that the increase in that pre-existing asymmetry as a result of the exchanges of information in question would not be sufficiently harmful to competition. Even supposing that such asymmetry of information existed, Crédit agricole’s argument comes up against the useful effect that must be guaranteed by the concept of ‘restriction by object’ and, more generally, by Article 101 TFEU.

As regards, in the third place, the applicants’ arguments that the conduct at issue is justified in the light of its pro-competitive effects, the Court points out that the pro-competitive effects alleged by the applicants need not, as such, be taken into consideration when classifying the conduct at issue as a ‘restriction by object’.

In any event, even supposing that the alleged ‘favourable’ effects of the conduct at issue could or should be taken into account, in one way or another, for the purposes of classifying it as a ‘restriction by object’, the applicants have not demonstrated the existence of any favourable implications such as to call into question the classification of that conduct as a ‘restriction by object’.

In so far as the applicants also presented the conduct at issue as ‘ancillary restraints’ on the performance of their function as a market makers in SSA bonds, the Court observes that the case-law relating to the exception of ancillary restraints on legitimate agreements is, in any event, not applicable in the present case, since the applicants had not shown that their activity as market makers would have been impossible in the absence of the infringing conduct.

In addition, the Court rejected the arguments that market makers in SSA bonds were systematically at an informational disadvantage compared with counterparties that did not have a permanent presence on the market, so that they had to compensate for that information deficit by seeking information from a number of sources.

It cannot be accepted that undertakings try to offset the effects of factual situations which they consider to be excessively unfavourable, such as possible asymmetries of risk existing between operators on a market, by collusive practices designed to correct those disadvantages. Such factual situations cannot justify an infringement of Article 101 TFEU, especially as the applicants were not acting on the secondary market in SSA bonds solely as market makers and were carrying on that activity voluntarily.

Determination of the amount of the fines imposed on the applicants

In determining the amount of the fines imposed on the applicants, the Commission essentially followed the method set out in the 2006 Guidelines. ( 4 ) However, as regards the calculation of the basic amounts, the Commission decided to use a proxy value instead of the value of sales provided for in point 13 of those guidelines. As a starting point for the calculation of that proxy value, the Commission has taken the annualised notional volumes and values of SSA bonds (‘the annualised notional amounts’) that the banks concerned traded during their individual periods of participation in the infringement at issue. Those annualised notional amounts were then multiplied by an adjustment factor that the Commission constructed using 33 representative categories of SSA bond issued by eight issuers.

In that context, the applicants complained in particular that the Commission had infringed the 2006 Guidelines by relying on a set of representative SSA bonds and not on data from their own transactions to calculate the adjustment factor and by using public data from the Bloomberg platform, which inflated the adjustment factor (‘the BGN data’).

Credit Suisse also criticised the Commission for overstating the proxy value by including hedging transactions in the notional amounts applied to it.

As a preliminary point, the Court finds that, although, by adopting the 2006 Guidelines, the Commission has limited itself in the exercise of the wide discretion it enjoys as regards the method of calculating fines, it is entitled to depart from them, provided that it gives reasons and justifies its choice to the requisite legal standard.

However, where the Commission departs from the 2006 Guidelines not in their entirety – as point 37 authorises it to do – but only, as in the present case, from point 13, it cannot depart from the guiding principles and underlying logic of those guidelines. Thus, in implementing the methodology it defines, it must, in particular, ensure that it takes account of the best available data, subject to the thorough review, in law and in fact, of the EU judicature.

In the light of those clarifications, the Court notes, in the first place, that, in the contested decision, the Commission stated its reasons and justified to the requisite legal standard its decision to depart from the methodology set out in point 13 of the 2006 Guidelines and to base its calculation of the basic amount on a proxy value for the value of sales which was arrived at by multiplying the annualised notional amounts of each of the banks concerned by an adjustment factor calculated on the basis of the sample of 33 categories of SSA bond.

In that context, the Court rejects the applicants’ arguments that the Commission should have adopted a methodology for calculating the adjustment factor based on their own transactions.

In that regard, the Court points out that a methodology based on the transaction data of the banks concerned would entail carrying out calculations of a much greater complexity than those already complex calculations carried out in the present case, even though the representative nature of the SSA bonds retained precisely guarantees that the data taken into consideration remain relevant for the calculation of the fine and make it possible to reflect the economic importance of the infringement at issue with the degree of precision required by the case-law. Such an alternative methodology would place a disproportionate administrative burden on the Commission.

In the second place, the Court rejects the applicants’ arguments that the BGN data used by the Commission were inadequate for the purposes of calculating the proxy value in that they inflated the adjustment factor.

After pointing out that it was for the Commission to ensure that it took account of the best available data, the Court noted that, in the contested decision, the Commission had set aside, with reasons, the arguments relied on by the banks concerned during the administrative procedure to contest the use of the BGN data. It follows that the applicants cannot confine themselves to arguing before the Court that the data used by the Commission suffer from one or more shortcomings but, on the contrary, must demonstrate that, within the framework of the methodology which that institution lawfully determined, there are in fact data better than those used by that institution and that those data are in fact available.

In finding that the applicants were unable to submit data better than those used by the Commission, the Court also rejects Credit Suisse’s criticism that the method of compiling the BGN data was unknown. On that point, the Court emphasises that the BGN data constituted reference data among traders, which are compiled by a third party to the proceedings on the basis of the prices of several traders. Accordingly, it cannot validly be argued that, on the ground that the manner in which they are compiled is partly unknown, such reference data cannot be used by the Commission, particularly where Credit Suisse has in no way referred to market platforms providing more accurate or more relevant information than the Bloomberg platform.

According to the Court, the Commission could not be criticised for having used data which did not reflect Credit Suisse’s situation in all respects, when Credit Suisse itself did not have accurate and sufficiently representative data and was therefore obliged to use a methodology based on alternative data which were necessarily less accurate, in order to reconstitute a proxy value.

In the third place, the Court rejects Credit Suisse’s complaint that the Commission overestimated the proxy value of the value of sales by including transactions relating to the purchase of liquid assets in the notional amounts applied to Credit Suisse.

On that point, the Court states that, in the context of the 2006 Guidelines, the concept of ‘proxy value’, like that of ‘value of sales’, is intended to take as the starting point for calculating the fine imposed on an undertaking an amount which reflects the economic importance of the infringement at issue and the weight of that undertaking in it. It follows that the determination of the proxy value involves taking into account all of the transactions carried out on the market affected by the infringement, for each of the undertakings that took part in the infringement at issue.

While the methodology adopted by the Commission admittedly results in the notional amount exchanged on a given cash-purchase transaction being taken into account both for the seller of the SSA bond concerned and for its purchaser where they both took part in the infringement at issue, that twofold taking into account stems from the very principles governing the determination of fines under the 2006 Guidelines in the specific context of the present case. Moreover, to exclude from the calculation of the proxy value part of the transactions that indisputably fall within the scope of the alleged cartel would have the effect of artificially minimising the economic importance of the infringement at issue, thereby undermining the objective of effectively prosecuting and punishing infringements of Article 101 TFEU.

Having thus validated the proxy values adopted in respect of the applicants, the Court finally rejected their arguments challenging the gravity multiplier that the Commission had applied to those values in accordance with points 20 to 23 of the 2006 Guidelines.

In view of the fact that this gravity multiplier had been set at 16% for all of the banks concerned, Crédit agricole argued in particular that the Commission should have used a lower individualised multiplier for it.

Since Crédit agricole referred, in support of that argument, to judgments of the Court of Justice prior to the date of publication of the 2006 Guidelines, the General Court began by noting that those judgments could not require the Commission to take into consideration factors other than the intrinsic seriousness of the infringement at issue when determining the gravity multiplier under the 2006 Guidelines.

In addition, while the Commission cannot disregard the principle of equal treatment when calculating fines imposed on the basis of Article 101 TFEU, it is clear both from point 22 of the 2006 Guidelines and from the relevant case-law that the gravity multiplier reflects, in principle, the gravity of the infringement at issue and not the relative gravity of the participation in that infringement of each of the undertakings concerned. It is in that sense that points 19 to 22 of the 2006 Guidelines envisage the determination of the gravity multiplier for the infringement at issue and not for each undertaking that took part in it. Thus, the assessment of individual circumstances is, in principle, not carried out as part of the assessment of the gravity of the infringement, namely when setting the basic amount of the fine, but as part of the adjustment of the basic amount according to attenuating and aggravating circumstances.

In noting, moreover, that a gravity multiplier of 16% for an infringement such as that found in the contested decision cannot be regarded as inappropriate or disproportionate, the Court dismisses the arguments put forward in that regard by Crédit agricole.

In the light of all of the foregoing, the Court dismisses Crédit Suisse’s action in its entirety. By contrast, it annuls the contested decision in relation to Crédit agricole in so far as, first, it finds that Crédit agricole participated in the infringement from 10 January 2013 to 24 March 2015, and not from 11 January 2013 to 24 March 2015, and, secondly, it sets the amount of the fine imposed on Crédit agricole at EUR 3993000. However, in the exercise of its unlimited jurisdiction, the Court maintains the amount of the fine imposed on Crédit agricole.