Credit agricole and Credit agricole Corporate and Investment Bank v Commission (Competition - Agreements, decisions and concerted practices - Suprasovereign bond, sovereign bond and agency bond sector denominated in United States dollars - Decision finding an infringement - Judgment) [2024] EUECJ T-386/21 (06 November 2024)

Judgment By their action under Article 263 TFEU, the applicants in Case T‑386/21, Crédit agricole SA and Crédit agricole Corporate and Investment Bank (together, ‘Crédit agricole’), seek, first, annulment of Commission Decision C(2021) 2871 final of 28 April 2021 relating to a proceeding under Article 101 TFEU and Article 53 of the EEA Agreement (Case AT.40346 - SSA Bonds) (‘the contested decision’), in so far as it concerns them, and, secondly, the reduction of the amount of the fine imposed on them in that decision. By their action under Article 263 TFEU, the applicants in Case T‑406/21, UBS Group AG, as successor in law to Credit Suisse Group AG, and Credit Suisse Securities (Europe) Ltd (together, ‘Credit Suisse’), seek, in essence, annulment of the contested decision in so far as it concerns them. I. Background to the dispute A. The SSA bond sector The conduct at issue concerns suprasovereign bonds, sovereign bonds and agency (sub-sovereign) bonds denominated in United States dollars (‘SSA bonds’). SSA bonds are debt instruments enabling their issuer to raise funds to finance certain expenditure or investments. A bond issuer may issue a bond to borrow money (the notional amount) from an investor (the holder) for a fixed term, which may be short (for example, 2 years) or long (for example, 10 or 30 years) and at a predefined fixed or floating rate of interest. The bondholder receives interest (the coupon) from the issuer at regular intervals, as well as repayment of the notional amount at the agreed maturity. SSA bonds are categorised according to the identity of their issuer. Suprasovereign bonds are issued by supranational institutions, such as the European Investment Bank (EIB) or the International Bank for Reconstruction and Development (IBRD). Sovereign bonds are issued by central governments and may be denominated in their national currency and subject to their national law, or denominated in a currency other than their national currency and subject to a law other than their national law. Finally, agency bonds are issued by subnational governments or by institutions implicitly or explicitly supported by public authorities, such as the Caisse d’Amortissement de la Dette Sociale (CADES, France) or the Kreditanstalt für Wiederaufbau (KfW, Credit Institution for Reconstruction, Germany). Those bonds can be issued in either domestic or foreign currency. SSA bonds are offered for sale for the first time by, or on behalf of, their issuer on the primary market. They are then traded over-the-counter between investors on the secondary market, without a central exchange. The conduct at issue in the contested decision does not concern the primary market, but only subsequent transactions carried out on the secondary market. The operation of the latter market, described in recitals 10 to 64 of the contested decision, is briefly set out below for the purposes of the present judgment. 1. Price of an SSA bond on the secondary market (a) Price formation on the secondary market Once a bond has been issued on the primary market, it is traded between investors on the secondary market. Between the time of issue and the maturity of the bond, the price of the bond changes, potentially deviating from its issue price in response to various factors, such as changes in the economic environment (for example, general changes in market interest rates) or factors specific to the issuer (for example, perceived risk, liquidity or availability of more recent issues). A bond is generally traded by reference to its yield to maturity, set as a percentage of its notional amount. That yield is the return an investor can obtain if he or she buys that bond today at a given price and holds it until maturity, and therefore depends on the coupon and any gain realised on an initial purchase below the notional amount or any loss realised on an initial purchase above the notional amount. On the secondary market, the price of a bond is generally determined by reference to the yield and is influenced by various factors such as the volumes available and the price and availability of comparable bonds. A change in market interest rates or a change in the issuer’s credit rating may also affect the price at which a bond is traded on the market. (b) Two-way prices offered by market makers A ‘market maker’ is an institution or individual willing to buy or sell financial products on the market on a general and continuous basis rather than transaction by transaction, at prices it determines itself. The role of a market maker is to provide a continuous quotation on bonds by quoting two-way prices - namely a bid price and an ask price - regardless of the state of the market, in order to facilitate trading in those bonds. By making volumes of bonds available and quoting them continuously, the market maker promotes market liquidity, namely the possibility for investors to buy or sell bonds quickly. Although there are no official market makers on the SSA bond market, certain financial institutions - such as the applicants in the present case - choose to assume the function of market maker, in that they quote two-way prices. When banks act on the secondary market, they try to generate income by capturing the difference between the bid price and the ask price. The difference between those two prices is called the ‘spread’, which constitutes the bank’s income when a purchase and resale transaction is considered in its entirety (‘the combined buy and sell transaction’). In the context of such a transaction, the ‘mid price’, or ‘mid’, which represents the mid-point between the bid price and the ask price, and which is expressed in cents (to two decimal places), is an important reference point, in that knowledge of it generally enables traders and investors to deduce the bid and ask prices. Median prices are not posted on brokers’ screens, nor are they communicated to investors and other counterparties. Finally, traders base their pricing and other strategic decisions on different sources of information. The bid and ask prices actually charged by a bank may be influenced by the identity of the other party (for example, a major investor for the bank) or by the volume involved in the buy or sell transaction. 2. Trading positions and trading risk After a bond purchase transaction, a trader is in a ‘long position’, in the sense that he or she holds the bond in his or her portfolio until he or she ‘closes’ that position, namely sells the bond again, if possible at a higher price than that at which he or she bought it. The trader may decide to hold the long position for a certain period, which represents a risk (or ‘exposure’). If the value of the bond falls, the trader will hold a bond in his or her portfolio that is worth less than the price at which he or she bought it. Conversely, after selling a bond, a trader may find him or herself in a ‘short position’ if he or she has sold a bond that he or she did not hold, typically when he or she has sold a larger volume of bonds than he or she did hold. He or she then has to buy the bond from another trader to close his or her position. Such a short position involves a potential gain, if the purchase price to obtain the bond is lower than the sale price obtained by the trader, but can lead to a potential loss, if the purchase price is higher than the sale price. A trader may decide to take a short position, for example, because he or she expects the price of the bond to fall (and thus to make a profit equal to the difference between his or her selling price and his or her buying price) or to meet a request from a major client. A trader may therefore have a long position or a short position in a given bond. A trader may decide to hold that position, subject to the ‘risk limits’ imposed by his or her employer. Alternatively, the trader may buy or sell bonds to end investors or through a broker in order to limit or offset his or her exposure. Such a way of reducing risk is a form of risk ‘hedging’, which can be achieved by buying or selling either the same bonds or other highly liquid bonds, such as United States Treasury bonds, with a similar coupon or maturity. 3. Exchanges between banks via brokers Brokers are financial intermediaries who facilitate transactions between banks and are generally remunerated by means of a commission, which is a percentage of the purchase or sale transaction. Brokers operate electronic platforms that display banks’ bids and offers in real time. Such platforms show the name, coupon and maturity date of a bond, the volume (namely the notional amount, usually expressed in millions of currency) of bids or offers, and sometimes the yield spread over a benchmark bond. The bid and ask prices displayed at a given time are the best available prices, namely the most competitive prices at that time. A bank trader who sees a bid or offer on a brokerage platform may decide to ‘hit’ the bid (by selling bonds) or ‘lift’ the offer (by buying bonds), respectively, in which case he or she accepts the trade on the terms displayed on the platform. On the main brokerage platforms, when that happens, the bid or ask price accepted for the transaction is visible to all traders, but the identity of the counterparties is not disclosed. A trader interested in an offer listed on a broker’s platform can call the broker to obtain further information, such as the availability of additional volumes. Traders often hold ongoing online discussions (chats) with several brokers on Bloomberg or other instant messaging platforms. If the trader is interested, a short negotiation will then take place outside the platform (by telephone or online chat), during which the broker will act as intermediary between the two parties with a view to reaching agreement on the final price and volume of the transaction. Brokers do not reveal the identity of the traders involved in a given transaction, which guarantees traders’ anonymity and prevents them from obtaining information about a trader’s position or intentions, and from using that information to their own advantage. Using a broker also guarantees reliable execution of trades. It is generally accepted that the benefits to traders of using a broker outweigh the brokerage fees. B. The proceedings giving rise to the contested decision Following an application submitted on 4 August 2015, the European Commission, on 4 December 2015, granted Deutsche Bank AG conditional immunity pursuant to point 8(a) of the Commission Notice on immunity from fines and reduction of fines in cartel cases (OJ 2006 C 298, p. 17). That bank continued to cooperate and, throughout the investigation, provided the Commission with additional information and evidence concerning the infringement at issue. On 4 December 2015, 6 September 2016, 15 March 2017 and 12 November 2019, the Commission sent requests for information pursuant to Article 18(2) of Council Regulation (EC) No 1/2003 of 16 December 2002 on the implementation of the rules on competition laid down in Articles [101 and 102 TFEU] (OJ 2003 L 1, p. 1) to Bank of America Corporation and Merrill Lynch International (together, ‘BofA’), Crédit agricole, Credit Suisse and other banks. Between 21 and 24 November 2016, the Commission carried out announced inspections at the premises of BofA and Credit Suisse in London (United Kingdom). On 20 and 21 December 2018, the Commission initiated proceedings with a view to adopting a decision pursuant to Chapter III of Regulation No 1/2003, in accordance with Article 2(1) and Article 10(1) of Commission Regulation (EC) No 773/2004 of 7 April 2004 relating to the conduct of proceedings by the Commission pursuant to Articles [101 and 102 TFEU] (OJ 2004 L 123, p. 18), and addressed a Statement of Objections to Deutsche Bank, Credit Suisse, Crédit agricole and BofA (‘the Statement of Objections’). All the addressees of the Statement of Objections were granted access to the Commission’s investigation file. In accordance with Article 10(2) of Regulation No 773/2004, they all submitted written observations on the Statement of Objections. They also presented their arguments at a hearing, which took place on 10 and 11 July 2019. On 6 November 2020, the Commission sent to each of the addressees of the Statement of Objections a letter containing further clarifications on the method proposed by the Commission to determine the proxy for the value of sales for the purpose of calculating the amount of their fine (‘the letter on fines’). In response, Deutsche Bank submitted its observations on 4 December 2020 and Crédit agricole, Credit Suisse and BofA did so on 8 January 2021. C. The contested decision 1. Establishment of the infringement at issue In the contested decision, the Commission referred to numerous discussions between traders of the banks concerned via persistent or non-persistent chat rooms on the internet and via bilateral discussions by electronic means or by telephone, which took place over a period between 19 January 2010 and 24 March 2015 (recitals 90 to 92 of the contested decision). In that context, the Commission considered that the general objective of the collaboration between the traders concerned was to obtain a competitive advantage vis-à-vis their clients and competitors by coordinating their prices and trading activities and by exchanging commercially sensitive market information relating to their trading activities, prices and trading and pricing strategies in the SSA bond sector (recitals 102 and 638 of the contested decision). The Commission concluded that the banks concerned had knowingly substituted practical cooperation between themselves for the risks of competition, to the detriment of other market players and their clients. According to the Commission, the series of discussions at issue can be grouped into five categories of agreements or concerted practices, which are closely related and partially overlapping (recitals 613, 638, 643 and 758 of the contested decision). As is clear from recitals 103 and 613 of the contested decision, those five categories consist respectively of: – first , by coordinating the prices offered to counterparties (‘category 1’); – secondly , by coordinating the prices communicated to the market in general (‘category 2’); – thirdly , exchanges of current or prospective commercially sensitive information on their trading activities and secondary market trading patterns (‘category 3’); – fourthly , by exchanging, confirming and aligning their trading and pricing strategies (‘category 4’); – fifthly , by coordinating their trading activities (‘category 5’). As regards the legal classification of the conduct at issue, the Commission concluded that it was contrary to Article 101(1) TFEU on the ground that it consisted of agreements or concerted practices having as their object the restriction or distortion of competition in the SSA bond sector in the European Economic Area (EEA), as is clear from Article 1 of the contested decision. In order to reach that conclusion, in the first place, the Commission chose to consider that the conduct at issue had all the characteristics of agreements or concerted practices within the meaning of Article 101(1) TFEU (see Section 5.2.1 and, in particular, recitals 620 and 621 of the contested decision). In the second place, the Commission considered that, in view of their content, their objectives and the economic or legal context in which they took place, that conduct had the object of restricting or distorting competition by concerting on prices, sharing clients, sharing markets and exchanging commercially sensitive information which reduced the uncertainty prevailing on the market. In that regard, the Commission also considered that the same conduct had also distorted competition by creating an asymmetry of information between market participants, giving the banks concerned a competitive advantage which was all the more significant because it concerned an over-the-counter market characterised by significant uncertainty (see Section 5.2.2 and, in particular, recitals 739, 741 and 742 of the contested decision). In the third place, the Commission considered that the conduct at issue constituted a single and continuous infringement, in that it formed part of an overall plan pursuing the same anticompetitive objective. In that regard, the Commission noted in particular that the conduct concerned a homogeneous product traded in a similar manner, that the mechanism put in place was the same, that the banks involved had generally remained the same and changed only when one of the traders participating in the conduct at issue moved from one bank to another (see Section 5.2.3.2 and, in particular, recitals 757 to 763 of the contested decision). In the fourth place, the conduct at issue concerned SSA bonds issued by authorities and institutions established in the EEA, and was the subject of a significant volume of transactions between Member States and between the European Union and the countries forming part of the EEA. Accordingly, that conduct was, in the Commission’s view, capable of affecting trade between Member States and between the contracting parties to the EEA Agreement (OJ 1994 L 1, p. 3) (see Section 5.2.4.2 of the contested decision). In the fifth place, the Commission considered that Article 101(3) TFEU and Article 53(3) of the EEA Agreement did not apply to the present case (see Section 5.3 of the contested decision). 2. The imposition of fines The Commission stated that it had determined the amount of the fines imposed in accordance with the Guidelines on the method of setting fines imposed pursuant to Article 23(2)(a) of Regulation No 1/2003 (OJ 2006 C 210, p. 2; ‘the Guidelines’) (see Section 8.2 of the contested decision). However, given the particularities of the financial sector and, in particular, the fact that financial products such as SSA bonds do not generate sales in the usual sense of the term, the Commission, rather than using the ‘value of sales’ linked to the trading of SSA bonds to determine the starting point for calculating the fine, considered it appropriate to determine a proxy for the value of sales (‘the proxy’), based, first, on the notional amount of SSA bonds traded during the respective periods of participation by the banks concerned in the infringement at issue and, secondly, on the differences between the bid and offer prices of those SSA bonds (see Section 8.2.3 of the contested decision). For the rest, the Commission determined the amount of the fine imposed on each bank concerned by establishing a basic amount, calculated on the basis of the gravity and duration of the infringement, and by adjusting that basic amount on the basis of parameters specific to each bank. 3. The operative part The operative part of the contested decision is worded as follows: ‘ Article 1 The following undertakings have infringed Article 101 [TFEU] and Article 53 of the EEA Agreement by participating, during the periods indicated, in a single and continuous infringement regarding … SSA bonds covering the entire EEA, which consisted of agreements and/or concerted practices that had the object of restricting and/or distorting competition in the sector of … SSA bonds: (a) DB Group Services (UK) Limited and Deutsche Bank AG participated from 19 January 2010 until 28 March 2014; during this period, Deutsche Securities Inc. participated from 22 March 2010 until 25 February 2013[;] (b) [BofA] first participated from 19 January 2010 until 23 October 2012; and again from 22 July 2014 until 27 January 2015; (c) [Crédit agricole] participated from 10 January 2013 until 24 March 2015; (d) [Credit Suisse] participated from 21 June 2010 until 24 March 2015; Article 2 For the infringement referred to in Article 1, the following fines are imposed: (a) [Deutsche Bank]: EUR 0 (b) [BofA]: EUR 12 642 000 (c) [Crédit agricole]: EUR 3 993 000 (d) [Credit Suisse]: EUR 11 859 000 … Article 3 The undertakings listed in Article 1 shall immediately bring the infringements referred to in that Article to an end in so far as they have not already done so. They shall refrain from repeating any act or conduct described in Article 1, and from any act or conduct having the same or similar object or effect. …’ II. Forms of order sought In Case T‑386/21, Crédit agricole claims that the Court should: – annul the contested decision, in particular Article 1 thereof, or, in the alternative, annul certain parts of that decision under Article 263 TFEU; – annul the fine or, in the alternative, reduce the amount of the fine, in particular reduce the fine to zero, under Article 261 TFEU; – order the Commission to take the necessary measures to comply with the judgment of the Court in accordance with Article 266 TFEU; – order the Commission to pay the costs. In Case T‑406/21, Credit Suisse claims that the Court should: – annul in whole or in part Article 1(d) of the contested decision; – in the alternative, annul in part Article 1(d) of that decision, in so far as it states that the price-fixing communications restrict competition by object or that Credit Suisse participated in a single and continuous infringement throughout the period specified in that article; – annul Article 2(d) of that decision; – in the alternative, annul in part Article 2(d) of the contested decision; – order the Commission to pay the costs or, in the alternative, an appropriate proportion thereof. In Case T‑386/21 and in Case T‑406/21, the Commission contends that the Court should: – dismiss the action in its entirety; – order the applicants to pay the costs. III. Law Having heard the parties, the Court decides to join Cases T‑386/21 and T‑406/21 for the purposes of the present judgment, in accordance with Article 68(1) of its Rules of Procedure. A. The requests that certain data be omitted from the public domain In the course of the proceedings, Crédit agricole, Credit Suisse and the Commission, in essence, requested that certain data be omitted from the public domain in the present judgment pursuant to Articles 66 and 66a of the Rules of Procedure. In that regard, it should be noted that, in reconciling the need to make judicial decisions public, on the one hand, and the right to protection of personal data and of business secrets, on the other, the court must seek, in the circumstances of each case, to find a fair balance, having regard also to the public’s right of access, in accordance with the principles set out in Article 15 TFEU, to judicial decisions (see judgment of 27 April 2022, Sieć Badawcza Łukasiewicz - Port Polski Ośrodek Rozwoju Technologii v Commission , T‑4/20, EU:T:2022:242 , paragraph 29 and the case-law cited). Furthermore, the confidential treatment of an item of information in the file is not justified in the case, for example, of information which is already public or to which the general public or certain specialist circles have access, information featuring also in other passages or documents in the case file in respect of which the party seeking to preserve the confidential nature of the information in question neglected to make a request to that effect, information which is not sufficiently specific or precise to reveal confidential data, or information which is largely apparent or may be deduced from other information which is legitimately available to the interested parties (see order of 14 March 2022, Bulgarian Energy Holding and Others v Commission , T‑136/19, EU:T:2022:149, paragraph 9 and the case-law cited). In addition, information which has been secret or confidential but which is five years old or more must, as a result of the passage of time, in principle be regarded as historical and therefore as having lost its secret or confidential nature unless, exceptionally, the party relying on that nature shows that, despite its age, that information still constitutes an essential element of its commercial position or that of interested third parties (see judgment of 19 June 2018, Baumeister , C‑15/16, EU:C:2018:464 , paragraph 54 and the case-law cited). In the light of the foregoing, the Court decides, in the present case, as regards the public version of the present judgment, to make anonymous the names of the natural persons mentioned in the contested decision. The Court also decides to omit certain data relating to the adjustment factor used by the Commission against Crédit agricole and Credit Suisse in calculating the amount of the fines imposed on them and which the Commission concealed in the contested decision both from the banks which took part in the infringement at issue and from the public. By contrast, the intelligibility of the reasoning of the present judgment, the passage of time, the stereotyped reasoning of the requests for the omission of certain data, the information already available to the public and the significant changes which occurred after March 2015 in the business of Crédit agricole and Credit Suisse linked to the SSA bond sector justify the Court disclosing to the public certain data relating to the criticisms of those banks and, in particular, to the methodology used by the Commission to calculate the fines imposed on those banks. Thus, first, the Court decides, in particular, not to conceal, in the public version of the present judgment, the information contained in Tables 2 to 4 of the annex to the contested decision communicated to Crédit agricole and certain information in Table 5 of that annex. Secondly, as regards Credit Suisse, the Court decides in particular not to conceal, in the public version of the present judgment, the information contained in Table 5 of the annex to the contested decision communicated to that bank, the average time separating its daily purchase and sale transactions from ‘partially balanced’ trades, the percentage represented by ‘partially balanced’ or ‘fully balanced’ trades in relation to the total number of trades carried out by that bank and the percentage represented by liquidity purchase trades in relation to the total number of transactions carried out by that bank. B. The purpose of the actions brought by Crédit agricole and Credit Suisse 1. Crédit agricole’s action First of all, it should be borne in mind that a decision adopted on the basis of Article 101 TFEU in respect of several undertakings, although drafted and published in the form of a single decision, must be regarded as a group of individual decisions establishing, in relation to each of the undertakings to which it is addressed, the breach or breaches which that undertaking has been found to have committed and, where appropriate, imposing on it a fine (judgment of 16 September 2013, Galp Energía España and Others v Commission , T‑462/07, not published, EU:T:2013:459, paragraph 90). It follows that, by bringing an action for annulment against such a decision, each addressee of that decision may bring before the EU Courts only those aspects of the decision which concern that addressee, while aspects concerning other addressees do not form part of the matter to be tried by the EU Courts (see, to that effect, judgment of 15 July 2015, Emesa-Trefilería and Industrias Galycas v Commission , T‑406/10, EU:T:2015:499, paragraph 117). However, that case-law does not deprive all of the addressees of such a decision of the possibility of challenging, in the context of their respective appeals, all of the factual elements on which the Commission relied in particular in concluding that the entirety of the conduct at issue should be classified as a ‘single and continuous infringement’ or a ‘restriction by object’, notwithstanding the fact that the addressee challenging them did not take part in the entirety of that conduct. Consequently, Crédit agricole is not entitled to seek the annulment of Articles 1 and 2 of the contested decision in their entirety. At most, it can seek the annulment of the parts of those articles that concern it. The application for annulment of Article 1 of the contested decision or the application for annulment of the fine must therefore be understood as referring solely to Article 1(c) and Article 2(c) of that decision. Next, by its first and second heads of claim, as interpreted in the preceding paragraph, Crédit agricole, in essence, asks the Court to annul Article 1(c) and Article 2(c) of the contested decision and, in the alternative, to reduce the amount of the fine provided for in the latter article. It is therefore necessary for the Court, under Article 263 TFEU, to assess the legality of the contested decision in so far as it concerns Crédit agricole - as regards, first, the finding of infringement made in respect of it in Article 1(c) of the contested decision and, secondly, the fine imposed on it in Article 2(c) of that decision - and to exercise the unlimited jurisdiction conferred on it under Article 261 TFEU and Article 31 of Regulation No 1/2003 as regards the amount of the fine imposed on that bank. Finally, by its third head of claim, Crédit agricole also asks the Court to order the Commission to take the measures necessary to comply with the judgment to be delivered by the Court, in accordance with Article 266 TFEU. In the context of the review of legality based on Article 263 TFEU, the Court has no jurisdiction to issue directions to the institutions, bodies, offices and agencies of the European Union, even where they concern the manner in which its judgments are to be complied with (orders of 22 September 2016, Gaki v Commission , C‑130/16 P, not published, EU:C:2016:731, paragraph 14, and of 19 July 2016, Trajektna luka Split v Commission , T‑169/16, not published, EU:T:2016:441, paragraph 13). The same applies in the context of the exercise by the Court of its unlimited jurisdiction. It follows that Crédit agricole’s third head of claim should be rejected as of now on the ground of lack of jurisdiction. 2. Credit Suisse’s action By its first, second, third and fourth heads of claim, formulated solely on the basis of Article 263 TFEU, Credit Suisse seeks, in essence, the annulment in whole or in part of Article 1(d) of the contested decision and the ‘annulment’ in whole or, in the alternative, in part of the fine imposed on it in Article 2(d) of that decision. In response to a measure of organisation of procedure, Credit Suisse expressly confirmed that it was not asking the Court to exercise its unlimited jurisdiction under Article 261 TFEU and Article 31 of Regulation No 1/2003 and that, as regards the fine imposed on it, it was confining itself to asking the Court, under Article 263 TFEU, to review the legality of certain elements of that fine and the method of calculation underlying it. Therefore, as regards Credit Suisse, the Court must assess only the legality of the contested decision under Article 263 TFEU, including in relation to the fine imposed by that decision. C. The purpose of the infringement found in the contested decision In the context of its first and second pleas in law, alleging errors in the classification of the conduct at issue as, respectively, a ‘restriction by object’ and a ‘single and continuous infringement’, Crédit agricole repeatedly claims that, by Article 1 of the contested decision, the Commission found a single and continuous infringement consisting of five ‘autonomous’ infringements, corresponding to the five categories of conduct referred to in recitals 103 and 613 of that decision (see paragraph 35 above). Crédit agricole concludes from this that, in the contested decision, the Commission had to establish that ‘each of the five categories of conduct alleged constitutes an infringement by object’ and that, ‘if one of the five categories of conduct is not established with regard to [it], the single and continuous infringement defined by the Commission must be set aside’. The Commission contests that reading of the contested decision, essentially on the ground that it found in it only one ‘single and continuous infringement’, without first or concomitantly identifying separate infringements corresponding to conduct falling within one or other of the five categories referred to in recitals 103 and 613 of that decision. It should be noted that the exact determination of the scope and number of infringements found in the contested decision is essential in order to assess the legality of that decision. It is also essential in order to determine the operative nature of certain arguments raised by Crédit agricole, in particular those alleging errors in the classification of certain forms of conduct in one or other of the categories referred to in recitals 103 and 613 of that decision, or errors in the classification of a single and continuous infringement and in the participation of that bank in the infringement as well as errors in the classification of one or other of those five categories as a ‘restriction by object’. In that regard, the Court of Justice has held that, while a set of conduct may be classified as a ‘single and continuous infringement’, it cannot be inferred from that that each of those forms of conduct must, in itself and taken in isolation, necessarily be classified as a separate infringement. That can be the case only if, in the decision concerned, the Commission has chosen to identify and classify as such each of those forms of conduct and then to provide evidence of the involvement of the undertaking concerned to which they are attributed (see, to that effect, judgment of 16 June 2022, Sony Corporation and Sony Electronics v Commission , C‑697/19 P, EU:C:2022:478 , paragraph 67). Moreover, in interpreting the decision concerned, the General Court must take care not to confuse the concept of ‘conduct’ with that of ‘infringement’, unless it commits an error of law (see, to that effect, judgment of 16 June 2022, Sony Corporation and Sony Electronics v Commission , C‑697/19 P, EU:C:2022:478 , paragraph 78). In the present case, it must therefore be determined whether, in the contested decision, the Commission found the existence of a single and continuous infringement with an anticompetitive object or, on the contrary, and as Crédit agricole maintains, the existence of a single and continuous infringement, consisting of five separate and autonomous infringements each having such an object. In that regard, it is true that, in the contested decision, the Commission first assessed the anticompetitive object of the conduct at issue or of the categories of conduct at issue identified in recitals 103 and 613 (recitals 622 to 749), and only then found the existence of a single and continuous infringement (recitals 750 to 828). In so doing, by assessing from the outset the anticompetitive object of the conduct at issue or of the categories of conduct at issue, the Commission defined the evidential regime of the infringement (see, to that effect, judgment of 30 January 2020, Generics (UK) and Others , C‑307/18, EU:C:2020:52 , paragraphs 63 and 104) even before it had defined the contours of that infringement. Such structuring of the contested decision is likely to suggest that the Commission intended to make a prior and separate assessment of the harmful effect on competition of each of the forms of conduct at issue or each of the categories of conduct at issue and, consequently, to find that there were as many restrictions by object as there were forms of conduct or categories of conduct identified, even before finding that there was a single and continuous infringement. However, a reading of the contested decision as a whole does not support such a conclusion. First, Article 1 of the contested decision merely states, with regard to Crédit agricole, that that bank ‘infringed Article 101 [TFEU] and Article 53 of the EEA Agreement by participating, during the periods indicated, in a single and continuous infringement regarding … SSA bonds covering the entire EEA, which consisted of agreements and/or concerted practices that had the object of restricting and/or distorting competition in the sector of … SSA bonds’. Thus, unlike Article 1 of the decision at issue in the case which gave rise to the judgment of 16 June 2022, Sony Corporation and Sony Electronics v Commission (C‑697/19 P, EU:C:2022:478 ), which referred to a single and continuous infringement composed of several separate infringements, Article 1 of the contested decision does not refer to any other infringement which may correspond to the conduct identified in the five categories of conduct referred to in recitals 103 and 613 of that decision and mentioned in paragraph 35 above. Secondly, that reading of the wording of Article 1 of the contested decision is confirmed by the grounds of that decision, in the light of which its operative part must be interpreted, if necessary (judgment of 15 May 1997, TWD v Commission , C‑355/95 P, EU:C:1997:241, paragraph 21). Thus, recital 827 of the contested decision states only that ‘the complex of collusive contacts between the parties constitutes one single and continuous infringement of Article 101(1) [TFEU] for which each of the parties is held liable’. Consequently, that recital does not contemplate that conduct falling within any of the five categories referred to in recitals 103 and 613 of the contested decision may also be classified, separately and autonomously, as infringements of Article 101(1) TFEU, a view confirmed, moreover, by recitals 757 to 827 of that decision. Furthermore, in recital 104 of the contested decision, the Commission expressly stated that the various categories of conduct it had selected were described for analytical purposes and that they were linked and often overlapped. Accordingly, Crédit agricole’s argument that the Commission found the existence of five ‘autonomous’ infringements corresponding to the five categories of conduct referred to in recitals 103 and 613 of the contested decision and qualifying as a ‘restriction by object’ is based on an incorrect reading of that decision. It must therefore be rejected. The same applies to Crédit agricole’s criticism that, in the present action, the Commission attempted to reinterpret its decision in order to have the Court find that it had not taken an independent view of the five categories of conduct referred to in recitals 103 and 613 of the contested decision. It also follows from that erroneous reading of the contested decision that the objections raised by Crédit agricole and based on errors of classification of the conduct at issue in one or other of the five categories of conduct referred to in recitals 103 and 613 of the contested decision, and their consequences for the classification of the conduct at issue as a ‘single and continuous infringement’ and as a ‘restriction by object’, or based on errors in the classification of one or other of those five categories of conduct as a ‘restriction by object’, are inoperative and must therefore be rejected. However, the arguments put forward by Crédit agricole against specific discussions will be taken into account at a later stage in order to determine whether the Commission was able, without committing an error, to classify the discussions at issue, taken as a whole, as a single and continuous infringement with an anticompetitive object. D. Admissibility of the criticisms directed against the interpretation of the discussions analysed in the contested decision In Section 4.2 of the contested decision and, more specifically, in recitals 112 to 577 thereof, the Commission made a chronological presentation of the events on the basis of the evidence in its possession. That presentation took the form of an analysis and interpretation of the transcripts of more than 120 discussions that took place between traders of the banks concerned between January 2010 and March 2015, mainly on chat rooms on the Bloomberg platform. It was on the basis of that analysis and interpretation that the Commission subsequently assessed the existence of agreements or concerted practices, the existence of a restriction of competition and the existence of a single and continuous infringement and its anticompetitive object. In the present case, the Commission contends, in essence, that certain criticisms made by Crédit agricole and Credit Suisse respectively are inadmissible in so far as they do not relate to its interpretation of the content of specific discussions mentioned in the contested decision, but to discussions which are envisaged in a general manner and which are not the subject of any specific criticism. Similarly, criticisms directed against discussions identified or addressed in annexes to the pleadings filed by those two banks are inadmissible. Finally, the factual elements and evidence not contested or not validly contested by the two banks should be taken as given. In that regard, it should be borne in mind that, under Article 21 of the Statute of the Court of Justice of the European Union and Article 76(d) of the Rules of Procedure, any application must state the subject matter of the proceedings and a summary of the pleas in law on which the application is based. According to well-established case-law, that indication must be sufficiently clear and precise to enable the defendant to prepare his or her defence or to enable the General Court to decide the case (see judgment of 25 January 2018, BSCA v Commission , T‑818/14, EU:T:2018:33 , paragraph 95 and the case-law cited). Thus, for an action, plea, complaint or argument brought before the Court to be admissible, the essential elements of fact and law on which the applicant relies must be apparent, at least summarily but in a coherent and comprehensible manner, from the text of the application itself. While it is true that the body of the application may be supported and supplemented, on specific points, by references to particular passages of documents annexed thereto, a general reference to other written documents, even if annexed to the application, cannot compensate for the absence of the essential elements of the legal argument which, by virtue of the provisions referred to in paragraph 93 above, must appear in the application (see, to that effect, judgment of 13 June 2013, Versalis v Commission , C‑511/11 P, EU:C:2013:386, paragraph 115). The annexes may therefore be taken into consideration only in so far as they support or supplement pleas in law, complaints or arguments expressly relied on by the applicant in the body of its pleadings and it is possible to determine precisely which elements they contain which support or supplement those pleas in law, complaints or arguments (see judgment of 9 September 2015, Samsung SDI and Others v Commission , T‑84/13, not published, EU:T:2015:611 , paragraph 33 and the case-law cited). Moreover, it is not for the Court to seek out and identify, in the annexes, the pleas in law, objections and arguments which it might consider to constitute the basis of the action, since the annexes have a purely evidential and instrumental function (see judgment of 17 September 2007, Microsoft v Commission , T‑201/04, EU:T:2007:289, paragraph 94 and the case-law cited). The principles set out in paragraphs 93 to 96 above therefore preclude undertakings, in the context of an action brought against a Commission decision adopted on the basis of Article 101 TFEU and in which that institution criticises undertakings for a specific number of identified forms of conduct, from confining themselves to challenging in a specific and substantiated manner only some of those forms of conduct in the body of their pleadings and, for the others, either allow the Court to infer from its findings in relation to those first forms of conduct what should be done about those other forms of conduct, even if they are identified or referred to in an annex, or refer the Court to an annex in which their criticisms are developed for the first time. Criticisms formulated exclusively in an annex and relating to conduct neither referred to nor, a fortiori, criticised in the body of an applicant’s pleadings cannot be regarded as arguments put forward by that party in the body of those pleadings. It is in the light of those considerations that the Commission’s objections to the criticisms made by Crédit agricole and Credit Suisse respectively must be examined. 1. The pleas of inadmissibility relating to Crédit agricole’s criticisms In its first plea in law concerning the classification of the conduct at issue as a ‘restriction by object’, Crédit agricole states that it disputes the anticompetitive object of all the discussions listed in the five categories referred to in paragraph 35 above. In that regard, Crédit agricole refers to 11 different discussions as ‘examples’, namely 1 discussion for category 1 (that of 12 March 2015), 4 discussions for category 2 (those of 31 January, 15 February, 21 March and 10 July 2013), 3 discussions for category 3 (those of 19 March, 24 May and 25 July 2013), 3 discussions for category 4 (those of 19 March and 3 June 2013 and that of 6 August 2014) and 3 discussions for category 5 (those of 18 January and 10 July 2013 and that of 12 March 2015). In its third plea in law, Crédit agricole also complains that the Commission wrongly assumed that it had knowledge of the information exchanged on the chat rooms concerned, for the purposes of classifying the conduct at issue as a ‘single and continuous infringement’. In support of its argument, the bank refers to three discussions - those of 10 and 31 January and that of 11 October 2013 - as ‘examples’. The Commission considers that Crédit agricole has not contested the majority of the other discussions of which it is accused, something that the bank refutes in its reply. To that end, the bank refers to its reply to the Statement of Objections, attached to the application. The Commission contends that that overall reference is inadmissible. In that respect, with regard to the first and third pleas in law raised by Crédit agricole, it follows from the considerations set out in paragraphs 93 to 98 above that, in the absence of specific and substantiated criticisms set out even summarily in the application, the criticisms that that bank made of the discussions other than those referred to in paragraphs 100 and 101 above and contained in the body of its written submissions must be declared inadmissible. Moreover, that defect in the application cannot be remedied by Crédit agricole’s general reference, in the reply, to its response to the Statement of Objections annexed to the application. Contrary to what Crédit agricole maintains, such an assessment does not infringe that bank’s right to an effective judicial remedy guaranteed by Article 47 of the Charter of Fundamental Rights of the European Union , on the ground, in essence, that it was obliged to contest only some of the discussions at issue because of the volume requirements applicable to its pleadings before the Court. It is clear from point 105 of the Practice Rules for the Implementation of the Rules of Procedure of the General Court that, in direct actions brought on the basis of Article 263 TFEU, the maximum number of pages for pleadings is set at 50 pages for the application and 25 pages for the reply. However, it should be noted that Crédit agricole’s application is less than 43 pages long and that the bank did not request that those maximums be exceeded, which is permitted under point 106 of those rules. It follows from the foregoing that the assessments made by the Commission in the contested decision as regards the discussions contested by Crédit agricole on the basis of general and unsubstantiated considerations or discussions contested by that bank on the basis of a reference to its reply to the Statement of Objections must be regarded as established and therefore as definitive (see, by analogy, judgment of 7 November 2019, Campine and Campine Recycling v Commission , T‑240/17, not published, EU:T:2019:778 , paragraph 105). By contrast, sufficiently specific and substantiated criticisms of the discussions referred to in paragraphs 100 and 101 above are admissible and will be analysed on the basis of the time stamps appearing in the transcripts of those discussions produced by Crédit agricole or by the Commission. 2. The pleas of inadmissibility relating to Credit Suisse’s criticisms In the second part of its first plea in law, Credit Suisse claims that the Commission infringed Article 101 TFEU by concluding that ‘price discovery’ restricted competition by object. In that context, it merely states that that expression refers to ‘communications compris[ing] information on (i) the true value of a bond, (ii) the trading strategies of counterparties, (iii) market makers’ inventory positions and (iv) order flows’. In response to a plea of inadmissibility in which the Commission argued that the second part of Credit Suisse’s first plea in law was inadmissible because the term ‘price discovery’ was imprecise, Credit Suisse annexed to the reply a table listing the 25 discussions which it described as ‘price discovery communications’. The table specifies the date of each of those discussions, the recital of the contested decision referring to it and its reference in the Commission’s file. In that regard, it follows from the considerations set out in paragraphs 93 to 98 above that, in the absence of substantiated and specific criticisms formulated even summarily in the application or even in the reply and its annex, the criticisms which that bank considers to have made of the Commission’s interpretation of the content of the 25 discussions referred to in paragraph 109 above must be declared inadmissible. It follows from the foregoing that the assessments made by the Commission in the contested decision as regards discussions contested by Credit Suisse on the basis of general and unsubstantiated considerations or discussions contested by that bank on the basis of a reference to an annex to the reply must be regarded as established and therefore as definitive (see, by analogy, judgment of 7 November 2019, Campine and Campine Recycling v Commission , T‑240/17, not published, EU:T:2019:778 , paragraph 105). By contrast, sufficiently specific and substantiated criticisms made in the context of other pleas in law or other arguments and directed against the Commission’s interpretations of the content of the discussions analysed in the contested decision are admissible and will be analysed on the basis of the time-stamp which appears in the transcripts of those discussions produced by Credit Suisse or by the Commission. That is the case with the criticisms relating to the anticompetitive nature of certain discussions and, more specifically, of the criticisms directed against the interpretation of the content of the discussions of 28 September 2010, 8 February 2012 and 10 January 2013, mentioned briefly in the first part of the first plea in law, and against the interpretation of the content of the discussions of 12 March 2015 and the email of 24 March 2015, mentioned in the fourth part of the second plea in law. That is also the case with the criticisms made in the second part of the second plea in law, which do not concern the anticompetitive nature of certain discussions, but which are directed against the assessments made in recital 808 of the contested decision concerning the extent of Credit Suisse’s knowledge of the anticompetitive conduct, on the basis of the content of the discussions of 25 July and 22 October 2013 and of 9 January 2014. E. Crédit agricole’s and Credit Suisse’s claims for annulment In support of their action, the applicants put forward pleas in law which overlap to a large extent and which must be assessed together, while taking care to respond to the specific arguments raised by each of them. Essentially, their pleas in law are based on three categories of criticism which should be examined in turn and which allege, first, errors in the classification of the conduct at issue as a ‘single and continuous infringement’ and the extent of their participation in that infringement (second pleas in law of Crédit agricole and Credit Suisse), secondly, errors in the classification of that single and continuous infringement as a ‘restriction by object’ (first plea in law of Crédit agricole and Credit Suisse), and, thirdly, errors in the determination of the amount of their respective fines (fourth plea in law of Crédit agricole and third plea in law of Credit Suisse). Before doing so, however, it is necessary to examine Crédit agricole’s third plea in law, alleging breach of the principle of the presumption of innocence, in that the Commission wrongly assumed that the traders of the banks concerned were aware of the information exchanged on the chat rooms simply because they were connected to them. 1. Crédit agricole’s third plea in law, alleging breach of the principle of the presumption of innocence By its third plea in law, Crédit agricole complains that the Commission assumed, for the purposes of classifying the conduct at issue under Article 101(1) TFEU, that ‘the traders [involved and, in particular, its own] were aware of every piece of information exchanged in a chat room when the trader concerned did not participate in the discussion group or in the exchanges’. In support of that plea in law, Crédit agricole relies on the judgment of 21 January 2016, Eturas and Others (C‑74/14, EU:C:2016:42 , paragraphs 45 and 50), according to which ‘the presumption of innocence precludes … the mere dispatch of [a] message constitut[ing] sufficient evidence to establish that its addressees ought to have been aware of its content’ and according to which, if it cannot be established that an undertaking was aware of a message, there must be ‘other objective and consistent indicia that it tacitly assented to an anticompetitive action’. In addition, that bank argues that the case-law relating to exchanges of information in the context of face-to-face meetings cannot be transposed to exchanges of information in the context of chat rooms. Consequently, in the present case, the Commission could not validly consider that the communication of information to ‘non-active participants’ on a chat room was sufficient to consider that the latter had received it and were aware of it, in particular for the reason, given in recital 587 of the contested decision, that they had the possibility of going back into past discussions. That was also due to the fact that the Crédit agricole trader was a member of more than 100 chat rooms and obtained information from a large number of other sources which appeared before him on six to eight screens. In particular, the Commission could not regard as directly incriminating the discussions of 10 and 31 January 2013 and of 11 October 2013, on the grounds that, in the case of the former, the Crédit agricole trader did not log on to the chat room in question until after the exchange of information retained by the Commission and, in the case of the latter, that, although he did log on to the chat room in question, he did not actively participate in the incriminating exchanges. Furthermore, the Commission has not shown that the Crédit agricole trader had an interest in seeking out the information concerned or acted on the basis of that information. In addition, ignorance of that information prevented the Crédit agricole trader from distancing himself from it publicly. The Commission considers that Crédit agricole’s arguments are unfounded and that the criticisms levelled at the discussions of 10 and 31 January 2013 and of 11 October 2013 are inoperative, since Crédit agricole has not contested the other discussions of which it is accused. At the outset, the Commission’s allegation that the discussions held on 10 and 31 January 2013 and on 11 October 2013 were irrelevant must be rejected. Although, as pointed out in paragraph 103 above, Crédit agricole cannot be regarded as having made an admissible criticism, in the context of its third plea in law, of discussions other than those three discussions, the fact remains that the absence of an admissible challenge to a majority of the conduct called into question by the Commission cannot lead to the automatic conclusion, without any individual analysis of the contested discussions, that the conduct found by the Commission, taken as a whole, was rightly classified as a single and continuous infringement with an anticompetitive object. As to the substance, it should be noted that the criticisms made by Crédit agricole raise two distinct issues. As regards, in the first place, the question whether the Commission may accept as evidence of anticompetitive conduct discussions which took place in the context of a chat room to which the undertaking concerned was connected, but in which it did not actively participate, it should be borne in mind that, according to settled case-law, passive modes of participation in an infringement, such as the presence of an undertaking in meetings at which agreements of an anticompetitive nature were concluded, without that undertaking clearly opposing them, are indicative of collusion capable of rendering the undertaking liable under Article 101(1) TFEU, since a party which tacitly approves an unlawful initiative, without publicly distancing itself from its content or reporting it to the administrative authorities, encourages the continuation of the infringement and compromises its discovery (see judgment of 22 October 2015, AC-Treuhand v Commission , C‑194/14 P, EU:C:2015:717 , paragraph 31 and the case-law cited). Admittedly, that case-law was developed in relation to meetings held in the presence of representatives of the undertakings concerned. However, there is no reason to believe that it should not be applied, by analogy, to discussions held in Internet chat rooms to which an undertaking is connected. Such a conclusion cannot be called into question by paragraphs 45 and 50 of the judgment of 21 January 2016, Eturas and Others (C‑74/14, EU:C:2016:42 ), referred to by Crédit agricole. The judgment of 21 January 2016, Eturas and Others (C‑74/14, EU:C:2016:42 ), did not concern, as in the present case, a persistent chat room in which the participants’ messages are delivered in real time to all the persons connected to it, who take note of them or, at the very least, can take note of them instantaneously. As is expressly clear from paragraph 7 of that judgment, the latter concerned a system operating in the same way as an electronic mailbox, where messages were delivered in the same way as electronic mail and, in order to be read, had to be opened in advance by the addressee, who therefore necessarily had to take an active approach in order to actually become aware of the content of the message addressed to him or her. Accordingly, the fact that the Court of Justice held, in essence, in the context of the judgment of 21 January 2016, Eturas and Others (C‑74/14, EU:C:2016:42 ), that knowledge of the content of a message sent by means of an instrument similar to an electronic mailbox could not be inferred from the mere fact that that message had been sent, cannot imply that, in the context of messages sent in a persistent chat room and which are delivered in real time to all the persons connected to it, the Commission must show that the message which is directed against an undertaking was actually read by the employee of that undertaking. Thus, in the present case, the Commission was entitled to consider that Crédit agricole had been aware of the discussions held on the persistent chat rooms, to which its trader was connected, even if the trader had not actively participated in those discussions or even if he had had at his disposal numerous other concomitant sources of information. That could only have been the case if Crédit agricole had demonstrated, by means of evidence that was certain and precisely time-stamped, that its trader had in fact not been aware of the offending message(s). Crédit agricole has not provided any such evidence, particularly with regard to the discussion that took place on 31 January 2013 between 11:02:08 and 11:05:38 on the PCHAT‑0x2000001313671 chat room, to which the trader from that bank had been connected since 7.16 on the same day. It may also be noted that that trader, although he did not take an active part in the discussion, reacted to the discussion in question at 11:29:08 in a message appearing four lines after the message of 11:05:38. The same applies to the discussion that took place on 11 October 2013 between 09:18:25 and 09:19:15 on the CHAT-fs:5257AB6D 02E00121 chat room, to which that bank’s trader had been connected since 08:44:58 on the same day. With regard, in the second place, to the separate question whether the Commission could accept, as evidence of anticompetitive conduct by Crédit agricole marking the beginning of its participation in the infringement at issue, the discussion of 10 January 2013, which took place in the context of a persistent chat room at a time when that bank’s trader had never before logged on using that bank’s identifiers, the Commission stated, in its observations following a measure of organisation of procedure, that it was not relying on the fact that, after logging on to the chat room, the trader from the same bank had become aware or had been able to become aware of the anticompetitive discussion that had taken place earlier that day, by going back through the history of the discussion. In that regard, the Commission stated the following: ‘Crédit agricole is liable for the entire single and continuous infringement for the entire period of its involvement, since it actively participated in all the conduct included in the single and continuous infringement … Even if the Commission were to demonstrate Crédit agricole’s knowledge ( quod non ), it is not required to show that [that bank] was aware of every piece of information exchanged, or that its trader actually read the entire exchange in the [chat room]. It is sufficient for the Commission to show that Crédit agricole was aware of the infringing conduct planned or implemented by the other participants in the cartel in pursuit of the common objective, or that it could reasonably have foreseen it and was prepared to take the risk.’ In the absence of exchanges of an anticompetitive nature that took place on the same day and after the Crédit agricole trader’s first connection to the chat room in question, that first connection alone is not sufficient to establish the starting point of that bank’s participation in the infringement at issue as 10 January 2013, even though, as a result of his previous employment relationship with BofA, the trader would have been aware that the exchanges that took place on that chat room were likely to be anticompetitive. First, it is clear from recitals 43 and 662 of the contested decision that the creation of chat rooms and participation in them is not, as such, anticompetitive. Accordingly, the Commission did not find against the participants the discussions that took place on such chat rooms, which had a social purpose or were intended to explore the possibility of bilateral transactions or to discuss the market colour, which was already known to all, even though those discussions took place during the infringement period. Thus, the anticompetitive nature of the discussions complained of by the Commission is linked solely to their purpose and not to the context in which they took place. Secondly, the discussion of 10 January 2013 is the first anticompetitive discussion held against Crédit agricole and marks, for the Commission, the beginning of that bank’s participation in the infringement at issue. In order to adopt 10 January 2013 as the date on which Crédit agricole began to participate in the infringement at issue, the Commission could not rely solely on the fact that that bank - through its trader - had logged on to the chat room at issue, in full knowledge of the fact that anticompetitive discussions had taken place on that chat room in the past and with the intention of continuing to take part in such discussions, as it implies in footnote 891 to the contested decision. It was still up to that institution to show that, on 10 January 2013 after Crédit agricole trader’s first connection to the chat room at issue, that bank had taken an active part or, at the very least, had passively assisted in an anticompetitive discussion, so that it had been able to transmit or obtain commercially sensitive information, of such a kind as either to influence the conduct on the market of an actual or potential competitor, or to disclose to such a competitor the conduct which Crédit agricole had decided to adopt on the market or which Crédit agricole was planning to adopt on that market (see, to that effect, judgment of 26 January 2017, Duravit and Others v Commission , C‑609/13 P, EU:C:2017:46 , paragraph 72). In the present case, neither the contested decision nor the documents before the General 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 first logged on to that chat room. Accordingly, the Commission could not, without committing an error and, in particular, without infringing the principle of the presumption of innocence applicable to proceedings for the implementation of Article 101 TFEU (see, to that effect, judgment of 21 January 2016, Eturas and Others , C‑74/14, EU:C:2016:42 , paragraph 38 and the case-law cited), find that, on 10 January 2013, Crédit agricole had taken part in anticompetitive conduct simply by logging on to the chat room at issue and consider that the starting point for that bank’s participation in the infringement had to be fixed at that date. In the light of the foregoing, Crédit agricole’s third plea in law should be upheld in so far as it concerns the discussion of 10 January 2013 and, as to the remainder, rejected as unfounded. However, the implications for the legality of the contested decision of the Commission’s error relating to that discussion of 10 January 2013 will be examined in paragraphs 550, 551 and 981 to 986 below, in the context of the duration of the infringement of which Crédit agricole is accused and in the context of the calculation of the fine imposed on that bank. 2. Crédit agricole’s second plea in law and Credit Suisse’s second plea in law, alleging errors in the classification of the conduct at issue as a single and continuous infringement (a) Preliminary observations In the event of a dispute as to whether there has been an infringement of the competition rules, it is incumbent on the Commission to provide evidence of the infringements it has found and to prove to the requisite legal standard, by a body of serious, specific and concordant evidence, the existence of facts constituting an infringement (see, to that effect, judgments of 31 March 1993, Ahlström Osakeyhtiö and Others v Commission , C‑89/85, C‑104/85, C‑114/85, C‑116/85, C‑117/85 and C‑125/85 to C‑129/85, EU:C:1993:120 , paragraphs 126 and 127; of 17 December 1998, Baustahlgewebe v Commission , C‑185/95 P, EU:C:1998:608 , paragraph 58; and of 6 January 2004, BAI and Commission v Bayer , C‑2/01 P and C‑3/01 P, EU:C:2004:2, paragraph 62). According to settled case-law, an infringement of Article 101(1) TFEU may result not only from an isolated act but also from a series of acts or even from continuous conduct, even though one or more elements of that series of acts or of that continuous conduct might also constitute, in themselves and taken in isolation, an infringement of that provision. Thus, where the various actions form part of an ‘overall plan’ because their identical object distorts competition within the common market, the Commission is entitled to impute responsibility for those actions on the basis of participation in the infringement considered as a whole (see judgment of 6 December 2012, Commission v Verhuizingen Coppens , C‑441/11 P, EU:C:2012:778 , paragraph 41 and the case-law cited). An undertaking which has participated in such a single and continuous infringement by engaging in conduct which was its own, which fell within the concepts of an agreement or concerted practice within the meaning of Article 101(1) TFEU and which was intended to contribute to the attainment of the infringement as a whole, may thus also be responsible for the conduct engaged in by other undertakings in the context of the same infringement for the entire period of its participation in that infringement. That is the case where it is established that that undertaking intended, through its own conduct, to contribute to the common objectives pursued by all the participants and that it was aware of the offending conduct planned or put into effect by other undertakings in pursuit of the same objectives or that it could reasonably have foreseen such conduct and was prepared to take the risk (see judgment of 6 December 2012, Commission v Verhuizingen Coppens , C‑441/11 P, EU:C:2012:778 , paragraph 42 and the case-law cited; see also, to that effect, judgment of 16 June 2011, Team Relocations and Others v Commission , T‑204/08 and T‑212/08, EU:T:2011:286, paragraphs 36 and 37). Thus, an undertaking may have participated directly in all the anticompetitive conduct making up the single and continuous infringement, in which case the Commission is entitled to impute to it responsibility for all that conduct and, consequently, for the infringement as a whole. Equally, the undertaking may have participated directly in only some of the forms of that conduct, but have been aware of all the other unlawful conduct planned or put into effect by the other participants in the cartel in pursuit of the same objectives, or could reasonably have foreseen that conduct and have been prepared to take the risk. In such cases, the Commission is also entitled to attribute liability to that undertaking in relation to all the forms of anticompetitive conduct comprising such an infringement and, accordingly, in relation to the infringement as a whole (judgments of 6 December 2012, Commission v Verhuizingen Coppens , C‑441/11 P, EU:C:2012:778 , paragraph 43, and of 24 June 2015, Fresh Del Monte Produce v Commission and Commission v Fresh Del Monte Produce , C‑293/13 P and C‑294/13 P, EU:C:2015:416 , paragraph 158). By contrast, if an undertaking has directly taken part in one or more of the forms of anticompetitive conduct comprising a single and continuous infringement, but it has not been shown that that undertaking intended, through its own conduct, to contribute to all the common objectives pursued by the other participants in the cartel and that it was aware of all the other offending conduct planned or put into effect by those other participants in pursuit of the same objectives, or that it could reasonably have foreseen all that conduct and was prepared to take the risk, the Commission is entitled to attribute to that undertaking liability only for the conduct in which it had participated directly and for the conduct planned or put into effect by the other participants, in pursuit of the same objectives as those pursued by the undertaking itself, where it has been shown that the undertaking was aware of that conduct or was able reasonably to foresee it and prepared to take the risk (judgments of 6 December 2012, Commission v Verhuizingen Coppens , C‑441/11 P, EU: C:2012:778, paragraph 44, and of 24 June 2015, Fresh Del Monte Produce v Commission and Commission v Fresh Del Monte Produce , C‑293/13 P and C‑294/13 P, EU:C:2015:416 , paragraph 159). It follows that the finding of the existence of a single and continuous infringement is distinct from the question whether liability for that infringement as a whole is attributable to an undertaking (see, to that effect, judgment of 26 September 2018, Infineon Technologies v Commission , C‑99/17 P, EU:C:2018:773 , paragraph 174). More specifically, as is clear from the case-law cited in paragraphs 148 to 151 above, three factors are decisive in determining whether an undertaking has participated in a single and continuous infringement. The first relates to the very existence of the single and continuous infringement, namely that the various forms of conduct at issue must form part of an ‘overall plan’ pursuing a single anticompetitive objective. The second and third elements concern the imputability of the single and continuous infringement to an undertaking. In that respect, first, that undertaking must have intended to contribute through its own conduct to the common objectives pursued by all the participants. Secondly, it must have been aware of the infringing conduct envisaged or implemented by other undertakings in pursuit of the same objectives, or must have been able reasonably to foresee it and been prepared to take the risk. In the present case, by its second plea in law, Crédit agricole claims that the Commission ‘has not provided adequate evidence or reasoning to show that it participated in a single and continuous infringement’. In the first part of its second plea, it argues that it ‘did not contribute to an overall plan’. In the second part of that plea, Crédit agricole argues that the infringement was not continuous. In that regard, first, in so far as Crédit agricole, in the sole heading of its second plea in law, criticises the Commission for not having provided an ‘adequate’ statement of reasons, it should be pointed out that the analysis of recitals 750 to 828 of the contested decision, which are devoted to demonstrating the existence of a single and continuous infringement in which that bank is alleged to have participated and which include, in particular, references to other recitals of that decision and replies to the arguments of the banks involved, enabled that bank to know the justification for the measure taken in order to defend its rights and enabled the Court to review the legality of the contested decision. In addition, the accuracy of the statement of reasons contained in the contested decision will be examined in the context of the response to the various arguments put forward by that bank and seeking to call into question the validity of the Commission’s assessments. Secondly, it follows from the wording of the first part of the second plea in law relied on by Crédit agricole that, in the context of that plea and as the Commission argued at the hearing, Crédit agricole does not contest the existence of an overall plan pursuing a single anticompetitive objective, but only its participation in that overall plan and therefore its participation in the single and continuous infringement found. As regards Credit Suisse, in the first and third parts of its second plea in law, that bank disputes the existence of a single and continuous infringement. In addition, in the second and fourth parts of that plea in law, the bank contests that it was responsible for the infringement for the period of time set out in the contested decision. In the light of the considerations set out in paragraphs 148 to 153 above, it is necessary to examine Crédit agricole’s and Credit Suisse’s arguments by distinguishing between, first, those directed against the finding of the existence of a single and continuous infringement and, secondly, those directed against the finding that liability for that infringement is attributable to them for a certain period of time. However, it is important to note that, in the context of their respective actions, Crédit agricole and Credit Suisse establish links between the questions raised by their second pleas in law and the arguments which they put forward in the context of their first pleas in law, alleging, in essence, an infringement of Article 101 TFEU committed by the Commission when it concluded that there was a restriction ‘by object’. Consequently, before examining the arguments referred to in paragraph 158 above, it is necessary to analyse certain findings of fact and certain assessments made by the Commission relating to the existence of agreements and/or concerted practices of an anticompetitive nature which are set out in recitals 593 to 621 of the contested decision. Those findings and assessments are likely to have an impact on the correctness of the Commission’s conclusions relating, first, to the existence of a single and continuous infringement and, secondly, to the imputability of responsibility for that infringement to Crédit agricole and Credit Suisse. (b) The existence of agreements and/or concerted practices of an anticompetitive nature As part of its assessment of the anticompetitive nature of the conduct it found, the Commission presented and interpreted the content of more than 100 discussions. However, Credit Suisse and Crédit agricole validly contested only some of the discussions concerning them, which means that the discussions that were not validly contested must be regarded as established and therefore as definitive, as must the Commission’s interpretation of them (see paragraphs 91 to 113 above). Thus, as regards Crédit agricole, it should be noted, in the light of the specific, precise and substantiated criticisms contained in its first plea in law, that that bank is entitled to challenge the anticompetitive nature only of the discussions of 18 and 31 January 2013, 15 February 2013, 19 and 21 March 2013, 24 May 2013, 3 June 2013, 10 and 25 July 2013, 6 August 2014 and 12 March 2015 (see paragraph 103 above). As regards Credit Suisse, it should be noted, in the light of the specific, precise and substantiated criticisms contained in the first part of its first plea in law and in the fourth part of its second plea in law, that that bank is entitled to challenge the anticompetitive nature only of the discussions of 28 September 2010, 8 February 2012, 10 January 2013 and 12 March 2015 and of the email sent by its trader to the Crédit agricole trader on 24 March 2015 (see paragraphs 112 and 113 above). Given that account must be taken of the anticompetitive nature of the conduct found by the Commission and, more specifically, of the interpretation of the content of the discussions analysed in the contested decision in order to examine the existence of a single and continuous infringement and the existence of a restriction by object, it is necessary, first, to recall the content of the discussions whose anticompetitive nature has been definitively established and, secondly, to assess the anticompetitive nature of the discussions that Crédit agricole and Credit Suisse are entitled to contest, given that those two banks do not contest either the existence of the discussions analysed in the contested decision or the content of those discussions and, in particular, the language used by the traders. (1) The discussions found to be anticompetitive In recital 613 of the contested decision, the Commission identified, for analytical purposes, the five categories of conduct referred to in paragraph 35 above, and illustrated the conduct at issue in particular in recitals 614 to 616 of that decision. In recital 614 of the contested decision, the Commission referred to numerous discussions in which participants expressed themselves in the following way about their prices: ‘gonna show the same … fck him’ (discussion of 1 February 2010 referred to in recital 127); ‘ok i will [s]how [him] the same’ (discussions of 8 February 2010 and 23 February 2012 referred to in recitals 129 and 377 respectively); ‘yeah cool … I am going 28/25 in libl’ (discussion of 24 February 2010 referred to in recital 131); ‘lets both bid same level’ (discussion of 10 March 2010 referred to in recital 143); ‘where shall we show[?]’ (discussion of 11 March 2010 referred to in recital 148); ‘I will bid the same’ (discussions of 25 March 2010 and 1 and 14 June 2011 referred to in recitals 157, 298 and 301 respectively); ‘where you want to bid/show?’ (discussions of 6 April 2010 and 25 April 2012 referred to in recitals 169 and 403 respectively); ‘lets both bid 41? … and split the trade?’ (discussion of 6 April 2010 referred to in recital 173); ‘being asked to offer 20mm … you might see it in a sec so lets [show] at the same level … 50?’ (discussion of 19 April 2010 referred to in recital 174); ‘ok cool will show same level … and wont improve’ (discussion of 2 June 2010 referred to in recital 190); ‘just seen the same effing enquiry … I'll bid 44 [too]’ (discussion of 12 August 2010 referred to in recital 215); ‘Let’s just keep that price up because I am not going to improve from there. Because they are just going to try and play one place against another’ (discussion of 13 October 2010, referred to in recital 236); ‘140 is fine man … will show that’ (discussion of 25 January 2011 referred to in recital 271); ‘will offer to miss’ (discussion of 11 February 2011 referred to in recital 277); ‘shall we switch prices etc at the same level?’ (discussion of 9 March 2011 referred to in recital 283); ‘in case he comes to [you] … maybe worth showing same level … so we can max the dough’ (discussion of 19 May 2011 referred to in recital 294); ‘tell him you see them like 20/17 or something’ (discussion of 15 July 2011 referred to in recital 309); ‘perfect … I’ll show the same’ (discussion of 6 October 2011 referred to in recital 337); ‘will show a worse price’ (discussion of 25 April 2012 referred to in recital 403); ‘lets leave it up there for 100mm?’, ‘then move it back a bps after that’ (discussion of 31 May 2012 referred to in recital 418). Other recitals in the contested decision also highlight the fact that participants expressed themselves in the following terms about their prices: ‘show them cheaper or whatever’ (telephone discussion of 25 September 2013 referred to in recital 528); ‘hey mate … I’m showing some 7-10yr ideas out to someone so was just that up. Where u wanna show out? 39 your level?’ (discussion of 5 February 2014 referred to in recital 539). In recital 615 of the contested decision, the Commission referred to numerous discussions in which participants exchanged information in the following terms: ‘just got an order from custy to sell 100mm germs … where do you see correct bidside’, ‘what the right bidside’, ‘where [you] got them marked?’ (discussions of 19 January 2010 referred to in recitals 116, 121 and 124 respectively); ‘where u buy cades 10/14 … i was aske those’ (discussion of 24 February 2010 referred to in recital 131); ‘being asked to offer 20mm eib 2.75 15 … where are these things at?’ (discussion of 6 April 2010 referred to in recital 165); ‘where you show? i have those as well’ (discussion of 19 May 2010 referred to in recital 185); ‘that’s what client is telling me they’ve seen away’ (discussion of 27 August 2010 referred to in recital 222); ‘where u gonna be bidding kfw 20’s if asked?’ (discussion of 31 August 2010 referred to in recital 228); ‘where you bidding the spain? seeing ti now as well’ (discussion of 7 July 2011 referred to in recital 304); ‘i think they must be sellers’ (discussion of 18 January 2012 referred to in recital 351); ‘is he the one that hit you[?]’ (discussion of 7 March 2012 referred to in recital 388); ‘where [you] guys marking these now?’ (discussion of 25 April 2012 referred to in recital 400); ‘oh yeah … that scumbag. Cool, thks for headsup’ (discussion of 28 August 2012 referred to in recital 434); ‘where would u bid that just out of interest’ (discussion of 12 July 2013 referred to in recital 507); ‘i showed 70, but he’s looking for 71’ (discussion of 14 March 2014 referred to in recital 547). Other recitals of the contested decision also highlight the fact that the participants exchanged information in the following terms: ‘what level these pups trading at [?]’ (discussion of 11 January 2013 referred to in recital 452 of the contested decision); ‘where you marking cades 04/17?’, ‘i am close to buying 200mm eib 06/18 … got a buyer already in the wings … and they were trading at lib+5 a short while back … still only at lib+18 … can easily come in another 5-7 bps … guy wants 33.5 … for 275mm … hes also thinking about selling me 250mm eib 03/18 … but wants 23.5 for them’, ‘that my bid in the cades 01/18 … in icap … and cantors … so going cheaper offer aint gonna help!’, ‘i just sold 73mm eib 03/20 in cantors … on a no post at 44.5’, ‘i just traded eib 03/20 at 44.5 … now they going up at 44 … and guy might even pay 43.5 … must have been a retail bu[y]er i guess’ (discussion of 2 July 2013 referred to in recital 495 of the contested decision); ‘we showed that switch out to them earlier … can’t believe they’re comping us … lol’, ‘i bid +12 and getting chiselled … showd nibs out at -1’ (discussion of 9 August 2013 referred to in recital 515 of the contested decision). Finally, in recital 616 of the contested decision, the Commission referred to numerous discussions in which the participants expressed themselves in the following way about their trading activities: ‘I’ll remove my offer’ (discussion of 24 February 2010 referred to in recital 131); ‘i’m gonna show this 215 bid a 200 offer if that doesn’t get in the way of what you’re doing’ (discussion of 12 August 2010 referred to in recital 211); ‘you want me to kill the bid?’ (discussion of 26 August 2010 referred to in recital 216); ‘ok I will show the same’ (discussion of 8 February 2010 referred to in recital 129); ‘where you want me to show’ (discussion of 8 March 2010 referred to in recital 139); ‘shall i kill me 144 offer’ (discussion of 24 September 2010 referred to in recital 230); ‘take it out for now man … don’t want [to] have to pay 15!’ (discussion of 20 October 2010 referred to in recital 243); ‘can i lift them first and cover my shorts?’, ‘no worries … let me know when you’re done’ (discussion of 9 November 2010 referred to in recital 248); ‘i’ll stay out of it for a while until u’re done’ (discussion of 30 November 2010 referred to in recital 260); ‘don’t worry man i’ll look after your posis while away’ (discussion of 13 December 2010 referred to in recital 262); ‘can show them tighter if it helps’ (discussion of 21 September 2011 referred to in recital 322); ‘i can kill it if u want’ (discussion of 2 November 2011 referred to in recital 341); ‘want me to show cheaper?’ (discussion of 19 January 2012 referred to in recital 356); ‘can u do me a favour and kill the bid if possible’ (discussion of 12 March 2012 referred to in recital 392); ‘i told him I’d sell once you’re out the way’ (discussion of 17 July 2012 referred to in recital 430); ‘can [you] kill that bid in the 08/15 pls’ (discussion of 28 August 2012 referred to in recital 434); ‘kill it for now if you can … just while i get the bid in’ (discussion of 15 October 2012 referred to in recital 440); ‘can you just stay out of it for the moment … as in don’t bid them up … and i will add your 5mm wherever i get mine back? … cool?’ (discussion of 23 January 2013 referred to in recital 461); ‘get yours down to a manageable position and then i’ll worry about mine’ (discussion of 13 March 2013 referred to in recital 475); ‘I can kill the offer if u like’ (discussion of 14 August 2013 referred to in recital 519); ‘let me kill offer’ (discussion of 4 March 2014 referred to in recital 541). Other recitals in the contested decision also highlight the fact that participants expressed themselves in the following terms about their trading activities: ‘[yo]u short kfw 10/22? i am short aswell. but told icap to give you all of them as i am not that keen to [co]ver them’ (discussion of 11 January 2013 referred to in recital 452); ‘sorry I didnt realised kbn 19 in tullets was you … i went better bid … can kill it if you wanty … we are short 10mm’ (discussion of 22 July 2014 referred to in recital 558). In the present case, first, Crédit agricole and Credit Suisse do not contest the Commission’s conclusion that those discussions established the existence of practices which had all the characteristics of an agreement and/or concerted practice within the meaning of Article 101(1) TFEU. Secondly, according to settled case-law, it is inherent in the Treaty provisions on competition that each economic operator must determine independently the policy which it intends to adopt on the internal market (see judgment of 23 November 2006, Asnef-Equifax and Administración del Estado , C‑238/05, EU:C:2006:734, paragraph 52 and the case-law cited). While that requirement of autonomy does not exclude the right of economic operators to adapt intelligently to the actual or expected conduct of their competitors, it does strictly preclude any direct or indirect contact between such operators which is likely either to influence the conduct on the market of an actual or potential competitor, or to reveal to such a competitor the conduct which it has decided to adopt on that market or which it intends to adopt on that market, where the object or effect of those contacts is to create conditions of competition which do not correspond to the normal conditions of the market in question, regard being had to the nature of the products or services offered, the size and number of the undertakings involved and the volume of that market (see judgment of 19 March 2015, Dole Food and Dole Fresh Fruit Europe v Commission , C‑286/13 P, EU:C:2015:184 , paragraph 120 and the case-law cited). First, the content of the discussions that took place between the traders of the banks concerned shows that they coordinated their prices (see paragraphs 166 and 167 above). As is clear from quotes such as ‘gonna show the same’; ‘ok I will [s]how the same’ or ‘lets both bid same level’, the participants agreed on the prices of their offers to specific clients as well as on the prices communicated to the market for a specific SSA bond. The content of the discussions that took place between the traders of the banks concerned also shows that they coordinated their trading activities (see paragraphs 170 and 171 above). In particular, as is clear from quotes such as ‘I’ll remove my offer’; ‘I’ll stay out for a while’ or ‘I can kill it if u want’, the traders agreed to refrain from bidding, to make an offer, or even to withdraw an offer from the market and, in particular, from a broker’s platform, when such an offer would have harmed one or other of those participants. That conduct could then lead to a split of the obligations concerned between the traders. The coordination of prices charged to specific clients or to the market and the coordination of trading activity therefore prevented the traders of the banks concerned from determining independently the policy they intended to follow on the market. It was particularly in the light of those considerations that the Commission rightly considered that the cooperation between the traders concerned and the respect they showed for each other’s interests were, at times, so close that they acted as if they were trading the same portfolio of bonds on behalf of one and the same undertaking. Secondly, it is apparent from the extracts reproduced in paragraphs 168 and 169 above and from other discussions examined chronologically in Section 4 of the contested decision that, during their discussions, traders disclosed, spontaneously or on request, commercially sensitive information which could be used by other traders to conduct their trading activities or to define their trading or pricing strategy. More specifically, the participants exchanged information on their future trading activities or strategy, such as, first, the transactions they had just completed, were in the process of completing or were considering completing, secondly, their personal quotation of the price of a bond (‘where do you see correct bidside’, ‘where [you] got them marked?’, ‘where you show?’), thirdly, their trading position, fourthly, the state of their inventory and, fifthly, the names of the clients who approached them, the requests made by those clients and the prices offered to those clients. The information exchanged was therefore specific to the banks involved, inaccessible to the public and clearly identifying, which made it sensitive. The sensitive nature of that information is corroborated in particular by a discussion of 15 November 2011, referred to in recital 343 of the contested decision, between the three main traders involved, namely, at the time, the Deutsche Bank trader, the BofA trader and the Credit Suisse trader. During that discussion, one of them explained: ‘this has to stay between us’. Moreover, those exchanges of information took place only within a restricted circle of traders who knew each other. Those exchanges enabled the traders in question to coordinate their conduct in terms of prices and trading conditions in an opportunistic manner and to avoid selling at a lower price or making a better offer than the other traders involved. Those exchanges thus created a favourable asymmetry of information, increasing transparency between the traders concerned and significantly reducing, to their benefit, the normal uncertainties inherent in the market, with the consequence of gaining an advantage vis-à-vis their clients and competing traders. Those exchanges therefore had a discriminatory dimension. Moreover, given their purpose, the information exchanged between the participants enabled each of them to have precise knowledge of the trading activities and pricing or negotiating strategies of their competitors with regard to a particular client or a specific obligation. Finally, the exchanges observed by the Commission concerned recent, current or future information, such as trades that had just taken place, that were in progress or that were envisaged in the short term. The exchanges of information found by the Commission in the present case were therefore not comparable to those at issue in the case giving rise to the judgment of 23 November 2006, Asnef-Equifax and Administración del Estado (C‑238/05, EU:C:2006:734), which led the Court of Justice to hold that a system for the exchange of information between financial institutions, such as a client credit information file containing anonymised information accessible in a non-discriminatory manner to all operators active on the market, did not, in principle, have the effect of restricting competition within the meaning of Article 101(1) TFEU. The exchanges of information at issue in the present case increased transparency between the participants in the infringement at issue. As a result, those exchanges reduced the uncertainty which existed on the market, a fact which Credit Suisse acknowledged at the hearing while maintaining that, in the atypical context of the SSA bond market, such exchanges did not have the typical effects of exchanges of current commercially sensitive information. The reduction in uncertainty caused by the exchanges of information found by the Commission benefited the banks in question, in particular, in that they were able to predict the conduct of other participants, to the detriment in particular of their clients and other competing traders. Consequently, the anticompetitive nature of the discussions, the interpretation of which is definitive and established, must be taken into account for the purposes of examining the second pleas put forward by Crédit agricole and Credit Suisse, relating to the existence of a single and continuous infringement, and for the purposes of examining the first pleas put forward by those banks, relating to the existence of a restriction by object. (2) The anticompetitive nature of the discussions, which Crédit agricole and Credit Suisse are entitled to contest It follows from the case-law that, where the Commission relies, in establishing an infringement of Article 101 TFEU, on documentary evidence, it is for the undertakings concerned not merely to put forward a plausible alternative to the Commission’s case, but to raise the insufficiency of the evidence relied on in the contested decision to establish the existence of the infringement. Similarly, where the Commission relies on direct evidence, the burden is on the undertakings concerned to show that such evidence is insufficient (see, to that effect, judgment of 8 September 2016, Goldfish and Others v Commission , T‑54/14, EU:T:2016:455 , paragraph 91 and the case-law cited). It is in the light of those considerations that the discussions that Crédit agricole and Credit Suisse are entitled to contest should be examined in turn. (i) The discussions that Crédit agricole is entitled to contest – The discussion of 18 January 2013 With regard to the discussion of 18 January 2013, classified in categories 3, 4 and 5 and referred to in recitals 456 to 460 and 674, 678, 686 and 746 of the contested decision, Crédit agricole does not put forward any concrete and specific argument relating to the content or to the anticompetitive nature of that discussion. By contrast, Crédit agricole argues that it only made a contribution unrelated to the offending exchanges. In that respect, it is clear from the transcript of the discussion in question, in which the traders from Deutsche Bank, Crédit agricole and Credit Suisse took part, that between 08:10:06 and 08:11:01 the Credit Suisse trader revealed certain trades that he had made or missed the previous night and, more specifically, the bonds and volumes that he had traded (‘saw some biz last night … seller of 50mm bng 17s, 25mm bng 23s and bought 50mm kfw 19s and sold 25mm … missed the bng 17s’). Furthermore, in response to a question from the Crédit agricole trader (‘nice what kind of a/c..all US?’), the Credit Suisse trader revealed the origin of the clients (‘US seller of bng, Asian buyer of kfw’). The Crédit agricole trader then expressed his thanks. Thus, the Crédit agricole trader, who entered the chat room at 07:12:08, received information from the Credit Suisse trader between 08:10:06 and 08:11:01 about recent transactions involving specific SSA bonds, their volume and the geographical origin of his clients. Crédit agricole has therefore not shown that the Commission erred when it considered, in recitals 456 to 460 and 746 of the contested decision, that the discussion of 18 January 2013 had enabled traders from Deutsche Bank, Crédit agricole and Credit Suisse to exchange information on transactions and clients and was therefore anticompetitive. A few minutes later, between 08:18:16 and 08:46:58, the Credit Suisse trader and the Deutsche Bank trader exchanged information, in particular about the bids they had made for a particular bond (BNG 17s). Credit Suisse then let the Deutsche Bank trader sell bonds (‘you wanna hit the bid?’, ‘nah, crack on man’). In that respect, the fact that the Crédit agricole trader did not take an active part in the exchange is not such as to deprive it of its anticompetitive character, which is moreover not disputed by that bank. Furthermore, Crédit agricole does not maintain, and there is no evidence in the file to suggest, that its trader left the chat room during the exchange between the Credit Suisse trader and the Deutsche Bank trader, that he was not present during the discussions between those two other traders or that he indicated to the other traders that he was taking part in the discussion from a different perspective. In the light of the case-law referred to in paragraph 126 above, the presence of the Crédit agricole trader during the second part of the discussion of 18 January 2013 is therefore evidence of complicity which is likely to give rise to liability under Article 101(1) TFEU. Consequently, Crédit agricole has not shown that the Commission erred when it found, in recitals 456 to 460 and 674, 678, 686 and 746 of the contested decision, that the discussion of 18 January 2013 had given rise to anticompetitive conduct on the part of the banks that took part in it. – The discussion of 31 January 2013 With regard to the discussion of 31 January 2013, classified in categories 2, 3 and 4 and referred to in recitals 464 to 467 and 655, 674 and 678 of the contested decision, Crédit agricole does not put forward any concrete and specific argument relating to the content of the exchanges between the Credit Suisse trader and the Deutsche Bank trader or to the anticompetitive nature of those exchanges. By contrast, Crédit agricole argues that it did not participate in those exchanges. In that regard, the transcript of the discussion in question shows that, between 11:02:08 and 11:05:38, the Deutsche Bank and Credit Suisse traders exchanged price information and coordinated their prices for a specific bond (eib 09/20) displayed on a broker’s screen (‘i just went 41 offered eib 09/20 … i am still short … but these look too tight i think … icap told me you were about to stick up a 41 offer aswell’, ‘i just pt it up at the same time … i’m flat, but same thing … i think they look wrong’, ‘they gone 47 bid … will go 42 offered … I am still short … but just see whats out there’, ‘cool … would offer there too … lets see what they say’). Moreover, it is clear from the transcript of the discussion in question that the Crédit agricole trader, who joined the chat room at 7.16 a.m., was present at the time of the exchanges in question and was able to become aware of the information exchanged. Moreover, the Crédit agricole trader reacted to that exchange of information at 11:29:08, that is to say, as the Commission pointed out in recital 467 of the contested decision, less than two minutes after the last remarks made by the Deutsche Bank trader about the bonds which had given rise to the conduct at issue. In addition, Crédit agricole’s argument that the Commission assumed that the bank was aware of the information exchanged during the discussion of 31 January 2013 must be rejected for the same reasons as those set out in paragraph 133 above. Consequently, Crédit agricole has not shown that the Commission erred when it found, in recitals 464 to 467, then 655, 674 and 678 of the contested decision, that the discussion of 31 January 2013 had enabled the Deutsche Bank and Credit Suisse traders to exchange commercially sensitive information and to coordinate their prices for EIB 09/20 bonds, in the presence of the Crédit agricole trader, and, consequently, that that discussion was anticompetitive. – The discussion of 15 February 2013 With regard to the discussion of 15 February 2013, classified in categories 2, 3 and 4 and referred to in recitals 468 to 472 and 655, 674, 678 and 746 of the contested decision, Crédit agricole argues that it did not give rise to coordination, in so far as it concerned a purchase of CADES 01/18 bonds by Deutsche Bank, which then resold half of them to Crédit agricole. In that regard, it is clear from the transcript of the discussion in question that, initially, the Deutsche Bank trader asked the Crédit agricole trader if the latter wanted to sell him CADES 01/18 bonds (‘you wann sell any cades 01/18?’). The Crédit agricole trader was open to selling, but alternatively offered to coordinate a joint purchase (‘u want me to sell mine? … or go with you on the buyside?’), which consisted of the Deutche Bank trader making an initial transaction and then selling part of the volume purchased back to the Crédit agricole trader (‘I’ll take whatever u don’t want’). Having agreed on the principle, the two traders then agreed, in writing and during a telephone conversation, on the price (‘ok do it’; ‘ok lifted 65’), the volume (‘50mm each … yeah?’) and the best time (‘lets wit a bit’; ‘ok’; ‘like 10mins or so’) to buy the bonds in question. First, as Crédit agricole maintains, a transaction did indeed take place between the Deutsche Bank and Crédit agricole traders following the discussion of 15 February 2013. However, that transaction followed, as the Commission found in recital 471 of the contested decision and as Crédit agricole essentially confirmed, a coordination of purchases between the two banks, leading to the two traders assisting each other in their activities and sharing the result between them. As a result of the discussion in question, the Crédit agricole trader did not submit a competing bid and the two traders thus acted as a single trader in order to purchase bonds at a price agreed between them. While such a transaction may prove to be in the interests of the two banks concerned, the fact remains that it cannot comply with the requirement that every economic operator must determine autonomously the policy which it intends to follow in the internal market (see judgment of 30 January 2020, Generics (UK) and Others , C‑307/18, EU:C:2020:52 , paragraph 78 and the case-law cited), as the Commission pointed out, in essence, in recital 472 of the contested decision. Secondly, in so far as Crédit agricole relies on a similar discussion of 23 May 2013 that was not held against a third-party bank, that argument cannot succeed. First, where an undertaking has, by its conduct, infringed Article 101 TFEU, it may not call into question the finding of such an infringement by relying on the fact that another economic operator was not the addressee of a decision finding the same infringement (see, to that effect, judgment of 25 March 2021, Xellia Pharmaceuticals and Alpharma v Commission , C‑611/16 P, EU:C:2021:245 , paragraph 166 and the case-law cited). Secondly, with the exception of a discussion on 23 May 2013 and a discussion on 30 May 2013 (see paragraph 242 below), Crédit agricole does not rely on any other evidence likely to show that its situation was comparable to that of the third-party bank referred to in paragraph 205 above, with regard to the single and continuous infringement at issue. Consequently, Crédit agricole has not shown that the Commission erred when it found, in recitals 468 to 472 and 655, 674, 678 and 746 of the contested decision, that the discussion of 15 February 2013 had enabled the Deutsche Bank and Crédit agricole traders to coordinate their negotiating strategy, including their price levels and the timing of their bids, and, consequently, that that discussion was anticompetitive. – The discussion of 19 March 2013 As regards the discussion of 19 March 2013, classified under categories 3 and 4 and referred to in recitals 477 to 479 and 674 and 678 of the contested decision, it is apparent from the transcript of that discussion that, in the space of a few hours, the Crédit agricole trader and the Credit Suisse trader discussed several bonds. That exchange concerned, first, the kbn/03/18s bond (‘just bid on kbn 03/18 … 25-50mm … bid +38 and missed’; ‘I just got hit in 10mm at 38 on the system … and then got hit at 38.5 in gfi’), secondly, the EIB 04/16s bond (‘I covered up some shorts today … was short 100mm of these eib 04/16 … bought most back … got some at 19 just now’) and, thirdly, the KFW 16s bond (‘decent purchase. I covered some kfw 16s … smallish at 17’). During the same exchange, the traders also disclosed the price at which they valued the Netherlands 09/15 bond (‘getting checked nther 09/15 … seller’; ‘where u marking 8/4?’; ‘i had them marked at 6’; ‘ok i showed 8/4 2 way indiction to the guy … not come back [yet]’) and the COE 04/17 bond (‘got hit in a few COE bits btw … coe 04/17 and coe 02/15 … in case u short’; ‘small long 4/17s … where u got them marked? … i like that bond … have them +53 bid here’; ‘got them 51 middle … so yeah same area’). The exchanges that took place on 19 March 2013 show that the traders exchanged information about their bid prices for certain bonds, the volumes involved, their inventory positions and the prices at which they valued a bond. The information exchanged concerned recently completed transactions and recent prices or a comparison of personal and current assessments of the price of a bond, which were not available to the public. There is no indication that the traders were exploring the possibility of concluding a transaction together. Moreover, with regard to one type of bond in particular (COE 04/17), the traders exchanged information on the prices of those bonds (‘have them +53 bid here’; ‘got them 51 middle’) despite the fact that they knew that a transaction between them was not conceivable on the grounds that each was in a long position and therefore already held such bonds (‘got hit in a few COE bits btw … coe 04/17 and coe 02/15 … in case u short’; ‘small long 4/17s’). It follows that those exchanges concerned commercially sensitive information and contributed to an increase in transparency between the traders and a reduction in the uncertainty prevailing on the market. Those exchanges therefore enabled the participants to adjust their respective strategies. Consequently, Crédit agricole has not shown that the Commission erred when it found, in recitals 477 to 479 and 674 and 678 of the contested decision, that the discussion of 19 March 2013 was anticompetitive on the ground that it had enabled the Deutsche Bank and Crédit agricole traders to exchange commercially current or prospective commercial sensitive information about their trading activities and to discuss their trading and pricing strategies. – The discussion of 21 March 2013 With regard to the discussion of 21 March 2013, classified under categories 2, 3 and 4 and referred to in recitals 480 to 487 and 674 and 678 of the contested decision, the Commission identified four exchanges that took place at different times of the day (between 08:08:22 and 08:18:40, between 11:29:22 and 12:26:04, between 14:43:55 and 14:51:28 and between 15:25:38 and 15:34:26). Crédit agricole maintains that the information exchanged during that discussion was legitimate. It also contests the Commission’s classification of the discussion of 21 March 2013 as category 2. On that point, it refers more specifically, by way of example, to the first and fourth exchanges. In that regard, first, it is important to stress that, contrary to what Crédit agricole’s argument suggests, it is not apparent from the contested decision that the Commission considered that the first, third and fourth exchanges, referred to in recitals 480, 484 and 485 of the contested decision, constituted evidence of coordination. It is clear from recitals 674 and 678 of the contested decision that the Commission considered that those three exchanges constituted exchanges of information of an anticompetitive nature. Secondly, as regards the second exchange, which took place on 21 March 2013 and is set out in recital 482 of the contested decision, it is apparent from recital 655 of that decision that the Commission classified that exchange as one of the cases of price coordination on the market in general. During that exchange, the Crédit agricole trader asked the Deutsche Bank trader whether a document that he had received through another channel was indeed that Deutsche Bank trader’s ‘axe sheet’ (‘is that [your] axe sheet?’), namely the list by which that trader provided information on the prices and quantities of bonds that he was willing to trade. The Deutsche Bank trader replied that the axe sheet was indeed his own. The two traders then discussed how Crédit agricole had obtained the document. The Crédit agricole trader said that it had been sent by mistake, that his supervisor had been annoyed and that it had caused a mess (‘he sent the sheet by mistake … what a mess’). The Deutsche Bank trader replied that he will speak to his sales desk (‘will speak to my sales’). The traders then talked about who had made the mistake and the Crédit agricole trader laughed and explained that ‘he’ shouldn’t have bothered as he knew the Deutsche Bank trader’s interests anyway (‘he shouldnt have bothered as i know your axes [anyway] lol’), to which the Deutsche Bank trader laughed and replied that that was true (‘lol … yeah exactly’). In that regard, it is true that that exchange does not reflect coordination on prices to show to the market generally falling within category 2, contrary to what is stated in recital 655 of the contested decision. The exchange shows that the axe sheet sent by the Deutsche Bank trader to Crédit agricole was the result of an error made by a Deutsche Bank employee. Nevertheless, the second exchange during the discussion of 21 March 2013 shows that a Deutsche Bank employee sent the axe sheet of one of that bank’s traders to another bank, Crédit agricole. Furthermore, that second exchange shows that the traders in question exchanged sensitive information with the result that the Crédit agricole trader knew, ‘in any event’, the interests of the Deutsche Bank trader. It follows that that second exchange supports the Commission’s assessment that the discussion of 21 March 2013 was anticompetitive. Thirdly, the error of assessment made by the Commission as regards the classification of the discussion of 21 March 2013 in category 2 cannot call into question the finding that that discussion, taken as a whole, was anticompetitive and that the Crédit agricole trader had participated in it. In the first exchange on 21 March 2013, the two traders exchanged information about the prices of transactions they had just carried out involving a KFW bond (‘just bought 35mm kfw 04/16 and sold 100 mm kfw 01/14’, ‘i am showing them at +2’, ‘sold them … 100.89’, ‘i sold at 100.9125’) and about the identity of a client (‘this is a [central bank] u delaing with right’). During the third exchange on the same day, the traders in question also exchanged, with regard to an EIB 03/20 bond, on their position (‘long 75mm’), prices (‘where u marking’, ‘38/35’, ‘i bid 41’) and the client of the Crédit agricole trader (‘[Central bank] seller?’, ‘no [Asset Manager]’). The fourth exchange concerned a CDC 09/13 bond and the traders exchanged information about their trading position (‘you need any cdc 09/13? … got 3mm’, ‘nah scrappy long myself’) and prices (‘where you marking these cdc?’, ‘100.40/50’, ‘perfect … sold my cdc … at 100.40 … in canotrs … made a 100.40’). Thus, in the light of the content of the first, third and fourth exchanges which took place on 21 March 2013, the Commission was entitled to conclude, in recital 487 of the contested decision, that, during that discussion, the Crédit agricole trader and the Deutsche Bank trader had exchanged information on recent transactions, the identity of certain clients, their current positions and the current price of the bonds, information which could help the traders to reduce uncertainty on the market and to assess their pricing intentions for the various bonds. Consequently, Crédit agricole has not shown that the Commission erred when it found, in recitals 480 to 487 and 674 and 678 of the contested decision, that the discussion of 21 March 2013 was anticompetitive on the ground that it had enabled the Deutsche Bank and Crédit agricole traders to exchange current or prospective commercially sensitive information about their trading activities and to discuss their trading and pricing strategies. – The discussion of 24 May 2013 As regards the discussion of 24 May 2013, classified under category 3 and referred to in recitals 488 to 491 and 674 of the contested decision, Crédit agricole argues, in essence, that it is not anticompetitive given the speed of the market. However, the transcript of that discussion shows that the Crédit agricole trader and the Credit Suisse trader exchanged commercially sensitive information about the bids they had made a few minutes earlier to a client known as ‘the sprayer’ for the EIB 08/16 bond (‘just bid eib 08/16 … 50mm … sprayer’; ‘ditto … i bid 99.80’; ‘s[a]me … he got .82 away’). The Commission was therefore right to consider that the traders had not coordinated their initial bids, but that they had exchanged recent or current information on their prices in the context of their respective interactions with the same client for the same bond. That exchange helped to increase transparency and therefore reduce uncertainty between the traders, with the result that the latter could take account of the information exchanged if the client in question were to return to them in the short term. Moreover, in recitals 659 and 716 of the contested decision, the Commission explained the reasons why information on recent transactions could be relevant for traders, and that assessment is confirmed by the explanations provided by Crédit agricole in point 2.15 of its application, according to which, ‘in order to be competitive, a trader needs to understand how prices have moved in the past in order to know what he needs to do to win business’. Consequently, Crédit agricole has not shown that the Commission erred when it considered, in recitals 488 to 491 and 674 of the contested decision, that the discussion of 24 May 2013 had given rise to an exchange of commercially sensitive information and, consequently, that that discussion was anticompetitive. – The discussion of 3 June 2013 With regard to the discussion of 3 June 2013, classified under category 4 and referred to in recitals 492 to 494 and 678 of the contested decision, Crédit agricole claims, in essence, that the Commission erroneously rejected its interpretation that the Deutsche Bank trader asked his trader what his valuation of a bond was, possibly with a view to setting up a ‘back-to-back negociation’, namely a negotiation whereby one of the two traders agrees to acquire bonds from a third party before reselling part of the bonds to the other trader. The Commission therefore wrongly ruled out the possibility that Deutsche Bank had sought to obtain liquidity. In that regard, it should be pointed out that the interpretation envisaged by Crédit agricole is presented by that bank itself as a mere ‘possibility’. Moreover, it is clear from the transcript of the discussion in question that, at the initiative of the Deutsche Bank trader (‘where you marking kfw 08/21?’), he and the Crédit agricole trader exchanged their personal assessment of the price of a bond and noted with satisfaction its similarity (‘sec … ct10-1/-5’, ‘yeah per[f]ect … i quoted +1/-4’). Finally, it emerged from that discussion that the Deutsche Bank trader was trading with a third party when he asked for the Crédit agricole trader’s valuation (‘not a regular guy’). No subsequent transaction with the Crédit agricole trader is foreseen in that discussion. By contrast, the Crédit agricole trader received information about the valuation of the Deutsche Bank trader which he could use if the third party in question decided to approach him. Consequently, Crédit agricole has not shown that the Commission erred when it considered, in recitals 492 to 494 and 678 of the contested decision, that the discussion of 3 June 2013 gave rise to an exchange of commercially sensitive information relating to the pricing of traders in the context of an ongoing negotiation, which resulted in increased transparency and reduced uncertainty on the market and, consequently, that that discussion was anticompetitive. – The discussion of 10 July 2013 With regard to the discussion of 10 July 2013, classified under categories 2, 3, 4 and 5 and referred to in recitals 501 to 506 and 655, 674, 678, 686 and 715 of the contested decision, Crédit agricole claims that that discussion admittedly led its trader to indicate to the Deutsche Bank trader that it was going to make an offer at the same price as that which it had previously communicated to him concerning an ASIA 10/18 bond. However, the Deutsche Bank trader did not maintain that offer and lowered his own, which rules out any coordination. Furthermore, that discussion demonstrates that Crédit agricole is a small market player whose conduct is not likely to have an impact on the market. In so doing, Crédit agricole criticises only the first part of the discussion referred to in recitals 501 to 503 of the contested decision (between 10:30:22 and 10:36:35), and not the second part referred to in recitals 504 to 506 (between 10:52:39 and 10:58:13). As regards the part of the discussion criticised, it should be noted that, before learning that the Deutsche Bank trader had lowered the price of his offer, the Crédit agricole trader asked the Deutsche Bank trader whether the price of the offer he was considering making on the basis of the information he had just received from the same trader bothered him (‘gonna go 23 offered? do u mind … dont want to do in your face’). The Deutsche Bank trader replied in the negative (‘nah its cool … i am 24 offered in cantors’). That discussion shows that the Crédit agricole trader requested and obtained authorisation from the Deutsche Bank trader to make an offer at a specific price for a given bond. Thus, the Crédit agricole trader did not take his decision to propose a price following an individual and autonomous assessment. The two traders determined together the price of the offer made by one of them. That discussion therefore reveals a coordination of the traders’ conduct on a subject that should have given rise to an individual assessment. Furthermore, and in any event, with regard to Crédit agricole’s argument based on the fact that the Deutsche Bank trader eventually lowered the price of his offer, the General Court has already had occasion to hold that sporadic and isolated cases of cheating or non-application of the cartel by a particular participant, especially where they concern a cartel of long duration, cannot in themselves demonstrate that that participant did not implement the cartel or adopted competitive conduct (judgment of 13 September 2013, Total Raffinage Marketing v Commission , T‑566/08, EU:T:2013:423, paragraph 254). Furthermore, the fact that Crédit agricole is only a small player on the SSA bond market does not call into question the anticompetitive nature of the discussion at issue. Finally, as regards the argument that the Commission did not treat a third-party bank in a similar manner in the light of a discussion on 30 May 2013 in which the latter participated, it must be rejected for the reasons adopted in paragraphs 206 and 207 above. Consequently, Crédit agricole has not shown that the Commission erred when it found, in recitals 501 to 506 and 655, 674, 678, 686 and 715 of the contested decision, that the discussion of 10 July 2013 had given rise to an exchange of commercially sensitive information and to price coordination and, consequently, that that discussion was anticompetitive despite the fact that the Deutsche Bank trader had ultimately lowered his offer. – The discussion of 25 July 2013 With regard to the discussion of 25 July 2013, classified in categories 3 and 4 and referred to in recitals 510 to 514 and 674 and 678 of the contested decision, Crédit agricole maintains that it does not concern forward-looking information. In that regard, it appears from the transcript of the discussion that, during the discussion, the traders from Crédit agricole and Deutsche Bank disclosed the volume of transactions they had carried out in relation to a new issue of a KFW bond that morning. The Crédit agricole trader reported that his offer to sell those bonds had been lifted by 50 million (‘got lifted in the new one this morn … 50 mm’). The Deutsche Bank trader replied that he had sold the same bonds for 500 million (‘i sold 500 mm of them this morning’). In addition, the Deutsche Bank trader revealed the type of buyer (‘asia’) to whom he had sold the said bonds for 500 million and the selling price to that buyer (‘+10’). The Crédit agricole and Deutsche Bank traders also discussed their current positions. In that regard, the Crédit agricole trader said: ‘had 100 on the book’, ‘still got 50’. The Deutsche Bank trader explained that, since the sale he had made that morning, he had managed to recover his short position after buying back bonds from various sources at a single price (‘i had zero … on the book’; ‘i lifted 14.5 everywhere … i got 150mm in creditex … on a no post … 100mm in icao … about 50mm in cantors … and rest in house’). A few minutes later, on the same day, the Deutsche Bank trader asked the Crédit agricole trader at what level he would price CADES 18 bonds (‘where are those cades 18? … seen anything in them?’). The two traders then revealed their respective current positions, namely that they were ‘flat’, namely in a neutral or zero position. Finally, in response to a further request from the Deutsche Bank trader (‘but i am mean where are they now’), the Crédit agricole trader gave his current assessment of the price of that bond (‘50/48 probs’). Therefore, during the discussion of 25 July 2013, the Crédit agricole trader and the Deutsche Bank trader exchanged information relating to the volume, price and origin of a client with regard to a recent transaction. In addition, on two occasions, the traders exchanged information on their inventory positions in relation to specific bonds. Finally, the said traders communicated on the current valuation, by one of them, of the price of a bond, namely the CADES 18 bond. Consequently, Crédit agricole has not shown that the Commission erred when it considered, in recitals 510 to 514 and 674 and 678 of the contested decision, that the discussion of 25 July 2013 had given rise to an exchange of commercially sensitive information relating to recent trading activities and to the participants’ current prices and positions and, therefore, that that discussion was anticompetitive. – The discussion of 6 August 2014 As regards the discussion of 6 August 2014, classified under category 4 and referred to in recitals 563, 564 and 678 of the contested decision, Crédit agricole claims, in essence, that the Commission erroneously rejected the idea that it could be a legitimate price request in the context of a search for liquidity. In that respect, it is clear from the transcript of that conversation that, during the conversation, the BofA trader asked the Crédit agricole trader how much he valued the KBN 03/18 bond (‘where [you] marking kbn 03.18?’). The latter said ‘48-45’, to which the BofA trader replied: ‘sounds about right’. Thus, during the discussion of 6 August 2014, the traders exchanged precise information about their own current assessment of the price of a bond, which was likely to reduce their uncertainty. Crédit agricole’s interpretation, according to which, in essence, it was a legitimate price request in the context of a search for liquidity, is not plausible and that bank does not show that the direct evidence available to the Commission regarding the discussion of 6 August 2014 was insufficient. First of all, nothing in the discussion that took place throughout the day of 6 August 2014, between 06:53:53 and 04:30:05, indicates that the BofA trader’s request would have been motivated by a desire to conclude a transaction with the Crédit agricole trader. Next, if the BofA trader had envisaged a transaction with the Crédit agricole trader, he would not have asked for an undifferentiated quote (‘where [you] marking kbn 03.18?’), but would have specified that his request was for an bid price or an offer price. Finally, as is apparent from the discussions referred to in recital 723 of the contested decision, the discussion of 6 August 2014 was preceded by other discussions of the same type, not validly contested by Crédit agricole, which show that that type of discussion took place in the context of a negotiation between one of the traders and a third party and was intended to verify that the valuation envisaged in the context of that negotiation was correct. Consequently, Crédit agricole has not shown that the Commission erred when it considered, in recitals 563, 564 and 678 of the contested decision, that the discussion of 6 August 2014 gave rise to an exchange of information on pricing strategy and, therefore, that that discussion was anticompetitive. – The discussion of 12 March 2015 With regard to the discussion of 12 March 2015, classified under categories 1, 3, 4 and 5 and referred to in recitals 573 to 575 and 674, 678, 686, 715, 746 and 848 of the contested decision, Crédit agricole admits that there was an exchange of information, but excludes that that discussion could have fallen within categories 1 and 5, as that discussion was ‘presumably’ intended to verify whether the Crédit agricole trader could acquire BNG 23 bonds from the Credit Suisse trader. As such an interpretation is plausible, the Commission could not therefore have qualified the discussion of 12 March 2015 as collusive and, therefore, as a ‘restriction by object’. According to the transcript of that discussion, the Crédit agricole trader asked the Credit Suisse trader whether he still had BNG 23 bonds (‘u still got bng 23’). The Credit Suisse trader replied in the affirmative, but emphasised that he had also been asked to make an offer (‘yeah … but being asked to price up too’). The Credit Suisse trader explained that he thought it was the same client (‘same guy i think’). At the same time, he sent the Crédit agricole trader an extract from the exchange he had just had with that client, which contained the volume and price offered to the client (‘14:14:36 50mm BNG 2.5 JAN23 @ ct7 +34.5’). The Crédit agricole trader reacted by saying that the client was making the rounds (‘man this dude gets aroudn lol’). He added that there was no problem and that he was leaving the client with the Credit Suisse trader (‘[no problem] man u crack on’). Forty minutes later, the Crédit agricole trader asked the Credit Suisse trader if he had completed the trade (‘u get lifted?’). Thus, first, that discussion clearly reveals an exchange of very precise and up-to-date information on the price and volume of an offer made to an identified client as part of an ongoing negotiation with that client. That discussion therefore gave rise to an exchange of commercially sensitive information and an exchange of negotiation and pricing strategy. Those exchanges increased transparency between the traders and reduced their uncertainty. Secondly, that discussion also reveals that, following that exchange of information, the Crédit agricole trader refrained from making an offer to the client who had contacted him and who had also contacted the Credit Suisse trader (‘[no problem] man u crack on’). The volume and price finally proposed by the Credit Suisse trader to the client in question, and the fact that only that trader made a proposal to the client in question, were not therefore the result of an independent choice by the two traders, but of coordination made possible by the discussion in question. That discussion led to the joint determination of the trader who was going to make the offer in question at a price known to the participants. Thus, the Crédit agricole trader decided not to make an offer to a specific client and left the client to the Credit Suisse trader. That interpretation is confirmed by the fact that, several tens of minutes later, the Crédit agricole trader asked his contact whether he had carried out the transaction. That interpretation, accepted by the Commission, is not contradicted by that proposed by Crédit agricole, according to which that discussion was in fact an attempt at a legitimate exchange which failed. According to the actual wording of Crédit agricole’s argument, the purpose of the discussion of 12 March 2015 was only ‘presumably’ intended to verify whether the Crédit agricole trader could acquire the SSA bonds concerned in order to have sufficient stock to then resell those SSA bonds to a third party, which he could only do if he was in a position to purchase those SSA bonds. Thus, Crédit agricole’s argument demonstrates that even that bank is not in a position to determine with certainty the truth of the alternative interpretation that it proposes. Moreover, the exchange between the traders went far beyond what was necessary to assess whether the Credit Suisse trader had liquidity and whether an alleged transaction between the two traders could take place. The Credit Suisse trader specified that it was the same client and provided the volume and price of the offer he had made to that client as part of an ongoing trade. Finally, given the content of the preceding exchanges, the message ‘[no problem] man u crack on’ cannot reasonably be understood as expressing an inability on the part of the Crédit agricole trader to submit an offer or to propose a better price than the Credit Suisse trader, a price of which he was not, moreover, supposed to be aware. The Crédit agricole trader’s refusal to make an offer occurred only a few seconds after the Credit Suisse trader disclosed his offer to the client in question. The Crédit agricole trader therefore did not seek, as would have been the case under normal conditions of competition, to find an alternative solution in order to be in a position to make an offer to the client concerned. By contrast, 40 minutes later, he asked the Credit Suisse trader whether he had carried out the transaction. Thus, the interpretation proposed by Crédit agricole is not plausible and, in any event, does not satisfy the standard of proof required to challenge the Commission’s interpretation, as required by the case-law referred to in paragraph 186 above. Consequently, Crédit agricole has not shown that the Commission erred when it considered, in recitals 573 to 575 and 674, 678, 686, 715, 746 and 848 of the contested decision, that the discussion of 12 March 2015 had given rise to an exchange of commercially sensitive information and to coordination of prices and trading activity and, therefore, that that discussion was anticompetitive. It follows that the arguments put forward by Crédit agricole to demonstrate that the discussions of 18 and 31 January 2013, 15 February 2013, 19 and 21 March 2013, 24 May 2013, 3 June 2013, 10 and 25 July 2013, 6 August 2014 and 12 March 2015 were not anticompetitive are rejected. (ii) The discussions that Credit Suisse is entitled to contest – The discussion of 28 September 2010 With regard to the discussion of 28 September 2010, classified under categories 3, 4 and 5 and referred to in recitals 232 to 235 and 674, 678, 686 and 746 of the contested decision, Credit Suisse claims that, during that discussion, its trader proposed to remove a price displayed on a broker’s screen in order to carry out, instead, a transaction with the Deutsche Bank trader at the same price as that displayed on the broker’s screen. In that regard, it should be noted that Credit Suisse is contesting only the first part of the discussion of 28 September 2010, namely the exchanges which took place between 05:46:31 and 05:48:17. It is not challenging the second part of the discussion, which took place between 11:23:57 and 11:29:54 and which, according to the Commission, consisted of pooling market intelligence on the trading preferences and future activity of a specific client known to the participants. During the first part of that discussion, the Deutsche Bank trader, as the Commission stated in recital 233 of the contested decision, ‘presumably’ asked the Credit Suisse trader whether he was a buyer or seller of Italian bonds (09/13) (‘wahts you better way ITALY 09/13’). In that context, the Deutsche Bank trader stated that he was trying to negotiate with a third party (‘trying to get a block out’). In response, the Credit Suisse trader explained that he had just displayed a two-way price on the BGC broker’s screen and asked the Deutsche Bank trader whether he should remove that price (‘cool … I’ve just gone 146/143 in bgc … shall i kill it?’). The Credit Suisse trader said he was working on a ‘switch trade’ with ‘he knew who’ (‘I’m working [on] an order to sell Italy 13s to buy Spain 12s with u know who’). Having received no response from the Deutsche Bank trader, the Credit Suisse trader asked again whether he should withdraw his price (‘so will need to buy at 146ish to print it … shall i kill price?’). The Deutsche Bank trader then replied, in essence, that it was not necessary, that he was not sure whether his potential client would come or not and that he would try his luck (‘nah its cool … not sure if this guy will sell or not … but trying him now’). Thus, first, it emerges from the transcript of the discussion in question that, during that discussion, the Credit Suisse trader proposed withdrawing his price from the market after learning that the Deutsche Bank trader was negotiating with a third party. That proposal would have resulted in not competing with the Deutsche Bank trader and would have increased the chances that the third party would negotiate with the Deutsche Bank trader. Secondly, there is nothing in the remarks exchanged to suggest, as Credit Suisse does, that its trader proposed to remove a price displayed on a broker’s screen in order instead to carry out a transaction with the Deutsche Bank trader at the same price as that displayed on the broker’s screen. In that regard, Credit Suisse’s interpretation that the Deutsche Bank trader sought to obtain liquidity from the Credit Suisse trader in order subsequently to execute a larger client order relating to that bond is not plausible. Furthermore, Credit Suisse has not demonstrated that the direct evidence available to the Commission regarding the discussion of 28 September 2010 was insufficient. The Credit Suisse trader told the Deutsche Bank trader the two-way price he had just displayed on the BGC broker’s screen and also announced that he was negotiating a sale of the Italian bonds in question to an identified third party (‘I’m working [on] an order to sell Italy 13s to buy Spain 12s with u know who’). Given that the Credit Suisse trader was already in the process of negotiating a sale of Italian bonds to a third party, it is implausible that the Credit Suisse trader’s proposal to remove his price from the broker’s screen was motivated by the intention to conclude a transaction with the Deutsche Bank trader. The most plausible interpretation is that the Credit Suisse trader proposed to withdraw his price from the BGC broker’s screen in order to limit the extent of possible competition and to avoid any risk that that price would interfere with the Deutsche Bank trader’s intention to proceed with a block sale to a third party. That interpretation is further corroborated by a discussion not specifically contested by Credit Suisse which took place a few days earlier, namely on 24 September 2010, during which the Deutsche Bank trader proposed to withdraw his offer from the Tullets broker’s screen in order not to interfere with the Credit Suisse trader’s trading strategy (see recital 230 of the contested decision). Consequently, Credit Suisse has not shown that the Commission erred when it found, in recitals 232 to 235 and 674, 678, 686 and 746 of the contested decision, that the discussion of 28 September 2010 gave rise to an exchange of commercially sensitive information, exchanges relating to trading and pricing strategies and coordination of trading activity and, therefore, that that discussion was anticompetitive. – The discussion of 8 February 2012 Credit Suisse argues that the exchange of information on market makers’ inventory positions makes it easier to find trading opportunities between traders. By way of example, it refers to the discussion of 8 February 2012, classified in categories 3 and 4 and referred to in recitals 367 to 376 and 674 and 678 of the contested decision. In that regard, it should be pointed out that Credit Suisse refers only to the first part of the discussion which took place between Deutsche Bank and BofA traders on 8 February 2012 between 02:58:05 and 03:10:51. Credit Suisse does not direct any specific argument against the other parts of that discussion, which took place later in the day and which, according to the Commission, gave rise to exchanges about client requests and pending transactions. With regard to the first part of the discussion that took place on 8 February 2012, the transcript of that discussion shows that the Deustche Bank trader and the BofA trader successively discussed transactions that they were in the process of carrying out or were considering carrying out with the same client in relation to several bonds. During those exchanges, the traders discussed ongoing negotiations concerning BNG 10/16 bonds (‘just about to lose 50mm bng 10/16’; ‘getting [check]ed on that too’; ‘just sold them … 50mm’). The BofA trader then revealed the price he had shown the client for the BNG 10/16 bonds (‘i showed +127 fyi’) and the Deutsche Bank trader replied that that was the price at which he had just concluded a transaction with that client (‘same … thats where i have been lifted’). In addition, the traders exchanged information on the client’s behaviour with regard to other bonds (‘same guy checking me on 100mm new neds’; ‘anyone got any ned 06/16’; ‘flat’; ‘i have some but about to get lifted’; ‘same guy?’; ‘y[ea]h’; ‘hes lifting all the cheap names … think he [might] lift me in 100mm coe17’). Thus, the first part of the discussion of 8 February 2012 did not consist solely of an exchange of information on the traders’ inventory positions. Traders exchanged information that was not public and related to the volumes of specific bonds they had sold or were considering selling to one and the same identified client. In addition, in relation to a particular type of bond, the traders exchanged information on the price at which they had sold or were considering selling that bond. The discussion in question therefore reflects running commentary on the traders’ activities which gave rise to an exchange of commercially sensitive information. Credit Suisse has therefore not shown that the Commission erred when it considered that the first part of the discussion of 8 February 2012 constituted a pooling of information which enabled traders to judge the intentions of an identified client and the current negotiations and, as a result, to act on the market with less uncertainty. Consequently, Credit Suisse has not shown that the Commission erred when it found, in recitals 367 to 376 and 674 and 678 of the contested decision, that the discussion of 28 September 2010 gave rise to an exchange of commercially sensitive information and to exchanges relating to negotiation and pricing strategies and, consequently, that that discussion was anticompetitive. – The discussion of 10 January 2013 During the discussion of 10 January 2013, classified under categories 3, 4 and 5 and referred to in recitals 446 to 450 and 674, 678 and 686 of the contested decision, the Credit Suisse trader asked whether the bid he was seeing (presumably on the screen of one or more brokers) for KFW bonds was that of the Deutsche Bank trader (‘is that your bid in KFW 4.375 18?’), to which the Deutsche Bank trader replied in the affirmative. The Credit Suisse trader then explained that he was negotiating with a third party for a ‘switch trade’ whereby he would buy those KFW bonds and, at the same time, sell another type of bond (‘I’m being checked on these vs the kommun’). The Deutsche Bank trader then replied that he was in a short position, that he needed those bonds and that he was prepared to pay whatever price (‘i am short 50mm … badly need them … will pay whatever’). He then asked the Credit Suisse trader, first, whether the latter wanted him to withdraw his bid visible on the broker’s screen (‘you want me to kill the bid?’) and, secondly, whether the latter could obtain the bonds for him (‘if you can get these for me will pay whatever man’). The Credit Suisse trader replied in the negative, saying that he would do his best, but that it could be a waste of time (‘nah, thats cool … i’ll do my best, but could just be time wasting’). That trader also pointed out that he did not do much business with that client and that he was not really sure whether he could close the deal (‘guy who we dont do anything with, so not really sure’). The Deutsche Bank trader then again offered to remove his bid from the broker’s screen (‘you want me to kill the bid?’), to which the Credit Suisse trader replied that the client was ‘not shopping it around’. Credit Suisse argues that, during that discussion, the Deutsche Bank trader offered to ‘kill’ his bid, as he hoped the Credit Suisse trader would trade for him and therefore did not need to cover his short position on the broker’s screen. The Credit Suisse trader’s response of ‘no, that’s cool … I’ll do my best, but it might be a waste of time’ would mean that the Deutsche Bank trader did not have to kill his offer, as the trade negotiation with the third party was not sufficiently advanced to allow him to rely on it and abandon the bond search on the broker’s screen. However, Credit Suisse’s interpretation is not plausible in the light of what was said during the discussion of 10 January 2013 and the bank has not shown that the direct evidence available to the Commission regarding that discussion was insufficient. The fact that the Credit Suisse trader did not ask for the Deutsche Bank trader’s offer to be withdrawn, on the ground that he did not think that the client was ‘looking for the best deal’, shows that the proposal to withdraw the offer was interpreted by the Credit Suisse trader as a desire on the part of the Deutsche Bank trader not to post a competing bid and thus not to interfere with the Credit Suisse trader’s ongoing negotiations. Moreover, an autonomous competitor who, like the Deutsche Bank trader, needed the bonds to such an extent that he was prepared to pay any price, would have sought to maximise his chances of finding a seller via multiple different channels and would therefore not have offered to withdraw his bid on the broker’s screen. The only plausible explanation for the proposal made by the Deutsche Bank trader to withdraw his bid is therefore the desire to help the Credit Suisse trader in his negotiations with the third party. In any event, it should be pointed out that, even though the Deutsche Bank trader did not ultimately withdraw his bid, the two traders showed a common desire to coordinate their conduct and then jointly decided that Deutsche Bank’s bid could be maintained on the broker’s screen. As the Commission points out, the fact that the Deutsche Bank trader might be interested in a subsequent transaction with the Credit Suisse trader if he concludes the transaction with the third party does not justify such concertation between competitors, since each of them has to make an independent competing offer in order to conclude a transaction with a third party. In addition, the discussion of 10 January 2013 gave rise to an exchange of commercially sensitive information about a potential transaction that could be of interest to both traders and, in particular, about the conduct of a client. The Deutsche Bank trader also communicated information about his situation. That information gave the Credit Suisse trader confidence that he would have a buyer who would be prepared to pay any amount for the bonds he would buy from his client, which the Credit Suisse trader could take into account in formulating his price to the client. Consequently, Credit Suisse has not shown that the Commission erred when it found, in recitals 446 to 450 and 674, 678 and 686 of the contested decision, that the discussion of 10 January 2013 gave rise to an exchange of commercially sensitive information, exchanges relating to negotiation and pricing strategies and to coordination of trading activity and, therefore, that that discussion was anticompetitive. – The discussion of 12 March 2015 With regard to the bilateral discussion that took place between the Credit Suisse trader and the Crédit agricole trader on 12 March 2015, also unsuccessfully contested by Crédit agricole and referred to in paragraphs 257 to 267 above, the arguments put forward by Credit Suisse are not likely to call into question the assessment that the said discussion was anticompetitive. First, the discussion that took place on 12 March 2015 does not merely reveal an exchange of very precise and up-to-date information about an offer proposed to a client in the context of an ongoing negotiation. That discussion also reveals a coordination of conduct that led the Crédit agricole trader not to submit an offer to a potential client who had contacted him, after consulting with the Credit Suisse trader (see paragraphs 259 to 261 above). Secondly, the alternative put forward by Credit Suisse that the Crédit agricole trader was trying to obtain liquidity from the Credit Suisse trader is not plausible, and Credit Suisse has not shown that the direct evidence available to the Commission was insufficient within the meaning of the case-law cited in paragraph 186 above. If, as Credit Suisse maintains, the Crédit agricole trader was trying to obtain liquidity, it would not have been useful for the two traders to exchange commercially sensitive information about the client’s strategy, the identity of the client and the price offer made by Credit Suisse to the client. It would have been sufficient for the Credit Suisse trader to explain to the Crédit agricole trader from the outset that he could not negotiate the bond with him or to provide him with a quotation. Consequently, Credit Suisse has not shown that the Commission erred when it considered, in recitals 573 to 575 and 674, 678, 686, 715, 746 and 848 of the contested decision, that the discussion of 12 March 2015 gave rise to an exchange of commercially sensitive information and to coordination of prices and trading activity and, consequently, that that discussion was anticompetitive. – The email of 24 March 2015 As regards the Credit Suisse trader’s email of 24 March 2015, to which the Crédit agricole trader was the addressee, classified under category 4 and referred to in recitals 576, 577 and 678 of the contested decision, Credit Suisse claims that it facilitates the identification of trading opportunities between traders, which is particularly important at the end of the month and quarter when market makers are supposed to flatten their portfolios. In that regard, it should be pointed out that the email in question was sent by the Credit Suisse trader to the Crédit agricole trader, in blind copy, and was entitled ‘CS US$ SSA BID AXES - month end: offer side bids’. That email contained an axe sheet, namely a list in which that trader provided information on the prices and quantities of bonds that he was willing to trade. It is apparent from recital 53 of the contested decision that an axe sheet shows the interest of a trader in a bank in buying or selling certain bonds on the market and is intended for the bank’s sales office or a broker. Thus, by sending such a list directly to a rival trader, the Credit Suisse trader informed a rival trader, namely the Crédit agricole trader, of the prices he was offering his clients for a series of bonds. Such information constitutes commercially sensitive information. That commercial sensitivity is not called into question by the fact that the list concerned only a limited number of bonds or that the prices were only indicative. Furthermore, Credit Suisse has not provided any evidence to show that its trader sought to use the axe sheet thus communicated to conclude a transaction with the Crédit agricole trader and thus to ‘flatten his portfolio’. Nor has Credit Suisse provided any evidence to show that the Crédit agricole trader sought to obtain liquidity from Credit Suisse for any of the 21 bonds mentioned in the list in question. The fact that such a list was not intended to be sent to a competing trader is confirmed by the discussion of 21 March 2013 between the Crédit agricole trader and the Deutsche Bank trader discussed in paragraphs 220 to 223 above. It is clear from that discussion that the sending of an axe sheet by a member of Deutsche Bank to Crédit agricole’s trader was a mistake and that such a mailing caused a ‘mess’ within Crédit agricole. Consequently, Credit Suisse has not shown that the Commission erred when it considered, in recitals 576, 577 and 678 of the contested decision, that the email of 24 March 2015 was anticompetitive. It follows that the arguments put forward by Credit Suisse to show that the discussions of 28 September 2010, 8 February 2012, 10 January 2013 and 12 March 2015 and the email of 24 March 2015 were not anticompetitive are rejected. (3) Conclusion on the discussions analysed in the contested decision It follows from all of the foregoing that Crédit agricole and Credit Suisse have not shown that the Commission erred when it considered that the discussions which it had analysed demonstrated the existence of agreements and/or concerted practices within the meaning of Article 101 TFEU of an anticompetitive nature and, more specifically, the existence, first, of anticompetitive coordination of prices and trading activity and, secondly, of exchanges of commercially sensitive information. The conclusion set out in paragraph 310 above does not, however, prejudge the response which must be given to the respective first pleas relied on by Crédit agricole and Credit Suisse and alleging, in essence, that the Commission infringed Article 101 TFEU by finding that the agreements or concerted practices found restricted competition ‘by object’. The fact that such conduct is anticompetitive does not necessarily mean that it automatically constitutes a restriction ‘by object’. In order to fall within the scope of a restriction of competition ‘by object’, the restriction must not only be anticompetitive in nature, but must reveal a sufficient degree of harmfulness to competition (see judgment of 30 January 2020, Generics (UK) and Others , C‑307/18, EU:C:2020:52 , paragraph 67 and the case-law cited). Crédit agricole’s and Credit Suisse’s arguments contesting the existence of a restriction ‘by object’ will therefore be examined later in the context of the reply to their respective first pleas in law. (c) The existence of a single and continuous infringement By the second part of its second plea in law, Crédit agricole challenges the continuing nature of the infringement found by the Commission. For its part, in its second plea in law, Credit Suisse alleges, in essence, lack of proof and inadequate statement of reasons, first, of the single nature of the infringement found by the Commission (first part) and, secondly, of the continuing nature of that infringement (third part). As noted in the minutes of the hearing, Credit Suisse withdrew its allegation of insufficient reasoning as to the single and continuous nature of the infringement at issue, initially raised in its second plea in law. The arguments of Crédit agricole and Credit Suisse must be examined by distinguishing, on the one hand, between those challenging the single nature of the infringement found by the Commission (first part of Credit Suisse’s second plea in law) and, on the other hand, those challenging its continuous nature (second part of Crédit agricole’s second plea in law and third part of Credit Suisse’s second plea in law). (1) The single nature of the infringement found In recitals 757 and 758 of the contested decision, the Commission considered that the series of agreements and/or concerted practices which it had previously identified formed part of an overall plan pursuing a common anticompetitive objective. For the banks concerned, that objective would have been to maximise their revenues, while limiting the losses that could result from the uncertainty linked to the conduct of other traders, by coordinating prices and trading activities and by exchanging commercially sensitive information relating to SSA bonds traded on the secondary market. The banks concerned would thus have sought to restrict or distort competition by substituting mutual assistance for competition between each other in order to improve their revenues from trading SSA bonds on that market, so that, at times, the situation would have been such that the traders of the banks concerned could have appeared to be trading on behalf of a single undertaking, rather than as competitors. Furthermore, in recitals 759 to 764 of the contested decision, the Commission set out, in essence, four ‘objective elements’ which, in its view, confirmed that the collusive conduct at issue was ‘linked and complementary in nature’. First, the products in question, namely SSA bonds, were homogeneous and traded in the same way (recital 760). Secondly, the mechanism used by the traders was the same, namely the use of constant communications in chat rooms reserved for them (recital 761). Thirdly, the undertakings concerned were the same, namely Deutsche Bank, BofA, Credit Suisse and Crédit agricole, and a change occurred only when one of the traders involved (who formed a close-knit group made up of the same people over time) changed employer (recital 762). Fourthly, the conduct at issue followed the same ‘pattern’. In particular, the traders of the banks concerned were in daily contact during certain periods and kept communication channels open throughout the day. They exchanged information on their current trading activities freely in the form of running commentary on events in progress, which enabled them to identify and exploit opportunities to coordinate their conduct (recitals 763 and 764). By the first part of its second plea in law, Credit Suisse claims that the Commission has not established the existence of an overall plan pursuing a single anticompetitive objective linking conduct before and after February 2013. Without there being any need to consider the exact scope of Credit Suisse’s challenge and, more specifically, whether that bank disputed the existence of an overall plan pursuing a single anticompetitive objective throughout the period covered by the infringement found by the Commission and, more specifically, before and during February 2013, which is not clear from the application, it is necessary to assess whether the Commission has demonstrated, first, the existence of such an overall plan before and during February 2013 and, secondly, the continuation of that overall plan after February 2013. (i) The existence of an overall plan pursuing a single anticompetitive objective before and during February 2013 As is clear from the discussions that took place before and during February 2013 and referred to in paragraphs 166, 168 to 171, 188 to 208 and 269 to 295 above, the Commission’s findings in the present case demonstrated that, during that period, the objective pursued by the conduct adopted by the participants, namely an exchange of commercially sensitive information and coordination of their pricing and trading activity, was anticompetitive. More specifically, it is clear from the Commission’s findings that the practices implemented by the traders of the banks concerned were aimed at exchanging commercially sensitive information and taking advantage of opportunities to coordinate their conduct when the opportunity arose and, in particular, when the participants showed a common interest in one or more specific SSA bonds or when they were approached by the same client concerning the same SSA bond. That conduct, which consisted in substituting mutual assistance for normal conditions of competition, led to increased transparency and therefore to a reduction in uncertainty and risk and, sometimes, to the management of the participants’ respective portfolios as if they had been a single portfolio. The participants were thus able to take advantage of their anticompetitive conduct vis-à-vis clients and competitors. The Commission therefore rightly considered that the sole anticompetitive objective pursued by the conduct adopted by the banks concerned was to maximise their revenues while limiting their losses. That objective is also reflected in certain explicit statements made during the discussions referred to in paragraphs 166, 168 and 170 above, such as ‘ok cool will show same level … and wont improve’ (discussion of 2 June 2010 referred to in recital 190); ‘where u gonna be bidding kfw 20’s if asked?’ (discussion of 31 August 2010 referred to in recital 228); ‘Let’s just keep that price up because I am not going to improve from there. Because they are just going to try and play one place against another’ (discussion of 13 October 2010 referred to in recital 236); ‘take it out for now man … don’t want [to] have to pay 15!’ (discussion of 20 October 2010 referred to in recital 243); ‘in case he comes to [you] … maybe worth showing same level … so we can max the dough’ (discussion of 19 May 2011 referred to in recital 294). Furthermore, the discussions that took place before and during February 2013 and referred to in paragraphs 166, 168 to 170, 188 to 208 and 269 to 295 above support the assessments made in the contested decision that, before and during February 2013, other objective elements confirmed that the anticompetitive conduct adopted by the participants was linked and complementary in nature and that, by interacting, that conduct contributed to the overall plan pursuing the anticompetitive objective described by the Commission. First, the anticompetitive conduct identified by the Commission concerned SSA bonds which, despite their variety, were traded in the same way and by the same traders at the same trading desks of the banks concerned. That assessment is illustrated by several discussions analysed by the Commission in the contested decision which show that, on the same day, the same traders exchanged information or coordinated their conduct in relation to different SSA bonds (see, for example, the discussion of 8 February 2012 referred to in paragraphs 279 to 286 above and the discussion of 18 January 2013 referred to in paragraphs 188 to 195 above). Secondly, the conduct at issue followed the same mechanism, namely, from the beginning of 2010, regular discussions within persistent multilateral chat rooms supplemented by bilateral discussions. Those discussions took place within a small, close-knit group of traders who exchanged views throughout the day on various SSA bonds (see, for example, the discussion of 28 September 2010 referred to in paragraphs 269 to 278 above and that of 31 January 2013 referred to in paragraphs 196 to 200 above). Thirdly, that conduct involved a stable group of undertakings, namely Deutsche Bank, BofA and Credit Suisse. That group of undertakings changed only when, in January 2013, the BofA trader changed employer and took up his post at Crédit agricole, leading to the involvement of the latter bank. Fourthly, the conduct at issue followed the same pattern, which took the form of the pooling of information or frank discussions about recent, current or future conduct that enabled the traders of the banks concerned to seize opportunities for coordination or exchanges of information and thus to pursue the single anticompetitive objective referred to in paragraph 317 above. The existence of an overall plan pursuing a single anticompetitive objective before and during February 2013 is also confirmed by other evidence in the file gathered by the Commission. During a discussion on 2 June 2010, referred to in recital 190 of the contested decision, the BofA trader informed the Deutsche Bank trader of the price at which he had sold bonds to a specific client. The Deutsche Bank trader subsequently offered the same price to the same client and told the BofA trader that he had sold the same bonds and that it had ‘worked out a dream’. Similarly, in recital 810 of the contested decision, the Commission referred to discussions from which it emerged that each of the three traders principally involved themselves recognised that they belonged to a discreet ‘club’ (‘telling you man … just like you, me and [Credit Suisse’s trader] have a chat room … all these accounts mst be in one chat room as well’; ‘we should defo include him in the club …’). Finally, on 10 January 2013, after leaving BofA and taking up a post at Crédit agricole, one of the traders involved in the conduct at issue examined by the Commission asked ‘to be added back in the house’, namely the persistent chat room of which the Deutsche Bank trader and the Credit Suisse trader were members. Thus, in so far as it had sufficient evidence to conclude that there was an overall plan pursuing a single anticompetitive objective before and during February 2013, it was not necessary, contrary to what Credit Suisse maintains, for the Commission to establish that the participants had explicitly and in advance subscribed to an overall agreement which defined their action on the market or, more generally, that they had formalised in advance their intention to engage in anticompetitive conduct. In the light of the foregoing and assuming that it intended to contest the existence of an overall plan between January 2010 and February 2013, Credit Suisse has not shown that the Commission erred when it considered that the conduct adopted by the traders of the banks concerned during that period formed part of the overall plan pursuing a single anticompetitive objective, as defined in the terms noted, in essence, in paragraph 317 above. (ii) The continuation of the overall plan pursuing a single anticompetitive objective after February 2013 As is apparent from recitals 94, 766, 767, 822, 823 and 847 of the contested decision, at the end of February 2013 Deutsche Bank prohibited its traders from using persistent multilateral chat rooms. Credit Suisse argues that that ban constitutes a ‘basic fact’ which made discussions on such chat rooms impossible, with the result that discussions between the traders of the banks concerned became less frequent, or even sporadic. As a result of that change in circumstances, the conduct engaged in by the traders after February 2013 would no longer have been based on the same ‘mechanism’ and would no longer have followed the same ‘pattern’. In those circumstances, it would no longer have been possible, after February 2013, to achieve the single anticompetitive objective defined by the Commission. The Commission accepts that there was a relative decline in the frequency of anticompetitive discussions as a result of the prohibition referred to in paragraph 337 above, but rejects Credit Suisse’s allegation that this decline was significant and sudden. In the present case, it should be emphasised that the prohibition imposed by Deutsche Bank on its traders from using persistent multilateral chat rooms is not liable to call into question the Commission’s findings concerning the existence of an overall plan pursuing a single anticompetitive objective. First, it is important to note that Credit Suisse does not dispute the objective factors referred to in paragraph 319 above, which relate, on the one hand, to the homogeneous nature of the products in question, namely SSA bonds, and, on the other, to the fact that the same banks concerned were involved, except where one of the participating traders changed banks. Secondly, the prohibition referred to in paragraph 337 above provoked a reaction from the Deutsche Bank traders, which is illustrated by a discussion which took place between two of that bank’s traders on 25 February 2013 and which is referred to in recitals 473 and 761 of the contested decision. During that discussion, a Deutsche Bank trader involved in the infringement at issue told one of his colleagues at the same bank that the ban would really hinder them (‘that’s gonna really hinder us’), as they would no longer be aware of the information exchanges between Crédit agricole and Credit Suisse traders (‘we not gonna know what flows are going on with [Crédit agricole’s trader] and [Credit Suisse’s trader]’). The other Deutsche Bank trader then explained: ‘what i do with [another trader] is we open a new room every day … as long as its not permanent chat, its ok’, to which his interlocutor, involved in the infringement at issue, replied: ‘ah ok … gonna have to do that then’. Thus, following the prohibition on the Deutsche Bank traders using persistent multilateral chat rooms, the traders of the banks involved in the infringement at issue replaced their discussions in persistent multilateral chat rooms with discussions in non-persistent bilateral chat rooms. In that regard, on the one hand, it is not disputed that, until August 2013, the Credit Suisse and Crédit agricole traders continued to discuss bilaterally on the same persistent multilateral chat room that had been used with the Deutsche Bank trader before and during February 2013, namely the PCHAT‑0x2000001313671 chat room known as ‘DB/CA/CS $ CHAT’. On the other hand, as is expressly clear from the discussion of 25 February 2013, quoted in paragraph 344 above, the prohibition on persistent multilateral chat rooms did not prevent the Deutsche Bank traders from using non-persistent bilateral chat rooms. Moreover, the traders of the banks concerned were accustomed to holding bilateral discussions as well, since they had held such discussions before and during February 2013, in addition to using the said multilateral chat rooms, which is not disputed. The traders of the banks concerned therefore sought to remedy the ban on the use of multilateral chat rooms imposed on the Deutsche Bank traders by means of a bilateral discussion network. That circumstance tends to demonstrate an intention to continue, after February 2013, the anticompetitive objective that had been pursued since January 2010. Thirdly, the discussions between the traders of the banks concerned did continue after February 2013. An examination of the post-February 2013 discussions shows that they were similar in anticompetitive content to the pre-February 2013 discussions. First of all, the traders continued to coordinate their prices, as illustrated, for example, by the extracts from bilateral discussions after February 2013 referred to in paragraph 167 above, by the bilateral discussions of 10 July 2013 (see paragraphs 236 to 243 above) and 12 March 2015 (see paragraphs 257 to 267 and 296 to 300 above), and by the email of 24 March 2015 (see paragraphs 301 to 308 above). Next, the traders continued to disclose, spontaneously or on request, information that could be used by other traders to conduct their trading activities or define their trading or pricing strategy, as illustrated, for example, by the extracts from bilateral discussions after February 2013 referred to in paragraph 168 above, as well as the bilateral discussions of 19 and 21 March 2013 (see paragraphs 209 to 216 and 217 to 227 above), 24 May 2013 (see paragraphs 228 to 232 above), 25 July 2013 (see paragraphs 244 to 248 above) and 12 March 2015 (see paragraphs 257 to 267 and 296 to 300 above). Finally, the traders continued to coordinate their trading activities, as illustrated, for example, by the extracts from bilateral discussions after February 2013, referred to in paragraphs 170 and 171 above, as well as the bilateral discussion of 10 July 2013 (see paragraphs 236 to 243 above) and that of 12 March 2015 (see paragraphs 257 to 268 and 296 to 300 above). Consequently, after February 2013, the traders of the banks concerned engaged in the same types of coordination, exchanged the same type of commercially sensitive information and pursued the same anticompetitive objective as before that date. Fourthly, Credit Suisse’s argument that the prohibition on the Deutsche Bank traders from using persistent multilateral chat rooms brought about a drastic change in the ‘mechanism’ and ‘pattern’ of the alleged single infringement does not call into question the finding that the anticompetitive objective pursued by the overall plan was maintained after February 2013. The single nature of an infringement results from the uniqueness of the objective pursued by the participants in the cartel and not from the way in which the cartel is applied (see, to that effect, judgments of 15 March 2000, Cimenteries CBR and Others v Commission , T‑25/95, T‑26/95, T‑30/95 to T‑32/95, T‑34/95 to T‑39/95, T‑42/95 to T‑46/95, T‑48/95, T‑50/95 to T‑65/95, T‑68/95 to T‑71/95, T‑87/95, T‑88/95, T‑103/95 and T‑104/95, EU:T:2000:77, paragraph 4127, and of 8 July 2008, Lafarge v Commission , T‑54/03, not published, EU:T:2008:255, paragraph 482). It has also been held that, since the objective of the anticompetitive practices remained the same, the fact that certain characteristics or the intensity of those practices had changed was not conclusive (see, to that effect, judgment of 24 March 2011, Aalberts Industries and Others v Commission , T‑385/06, EU:T:2011:114, paragraph 105). Furthermore, the single anticompetitive objective identified in the contested decision did not, contrary to what Credit Suisse suggests, depend on the organisation of daily discussions. As is clear from recitals 607 to 611 and 763 of the contested decision, the traders of the banks concerned pursued that objective by freely exchanging information on their current trading activities. Those exchanges took the form of running commentary on events as they occurred, enabling them to identify and exploit opportunities for coordination. In that way, those traders sought to achieve that objective in the context of anticompetitive discussions relating to particular negotiations or, at most, relating to information relevant to particular periods. It was therefore not necessary to have ongoing or daily discussions in order to pursue the anticompetitive objective in question. Moreover, even in the period prior to February 2013, anticompetitive discussions did not necessarily take place on a daily basis, as can be seen from the chronology of events set out in Section 4.2 of the contested decision. Finally, it is true that the ban on persistent multilateral chat rooms may have made it more difficult to identify opportunities for the traders from the banks concerned to coordinate or exchange commercially sensitive information, particularly in that it was more difficult for them to maintain ‘running commentary’. However, Credit Suisse does not dispute that the non-persistent bilateral chat rooms operated in the same way as the persistent multilateral chat rooms. Like the persistent multilateral chat rooms, the non-persistent bilateral chat rooms allowed connected traders to exchange information by sending instant messages. Moreover, like the persistent multilateral chat rooms, the non-persistent chat rooms allowed connected traders to comment, throughout the day and in real time, on events that were taking place. In reality, the difference between those two types of chat room lay essentially in the number of participants and whether or not it was necessary to open a new chat room each day. Consequently, without there being any need to assess the other justifications put forward by the Commission in the contested decision in response to the arguments of the banks involved to explain a relative decrease in the frequency of anticompetitive discussions, Credit Suisse’s argument that, after February 2013, it was no longer possible to achieve the single anticompetitive objective set out in the contested decision cannot be accepted. The first part of the second plea in law put forward by Credit Suisse is therefore rejected. In so far as, for its part, Crédit agricole challenges, in an incidental and scattered manner, the single nature of the infringement at issue, that bank’s allegations must also be rejected. Crédit agricole has no basis for arguing, at the end of its first plea in law and in the introductory part of its second plea, that the classification as a ‘single and continuous infringement’ would be incorrect if the Court were to find that the Commission had wrongly classified one of the five separate agreements or concerted practices as an infringement ‘by object’. That argument is based on the erroneous premiss that the Commission identified five separate infringements in the contested decision (see paragraphs 70 to 89 above). Nor is Crédit agricole entitled to suggest, in the context of the first part of its second plea in law, alleging a lack of contribution to an overall plan, that the single anticompetitive objective identified by the Commission for the single and continuous infringement at issue was defined too broadly, since any bank would seek to reduce its losses. While it is true that the single anticompetitive objective set out by the Commission includes the objective of maximising revenues and minimising losses by participating in the infringement at issue, it is clear from paragraph 317 above that the Commission also specified that the banks concerned had sought to achieve that objective by coordinating their conduct as regards prices and their trading activities and by exchanging commercially sensitive information relating to SSA bonds traded on the secondary market. Thus, the banks involved had sought, through mutual assistance and therefore by means of practices which did not fall within normal conditions of competition, to maximise their revenues and mitigate their losses to a greater extent than the banks which were not involved in the conduct at issue. (2) The continuous nature of the infringement found In recital 789 of the contested decision, under the heading ‘Continuous nature of the infringement’, the Commission considered that the conduct it had described in Section 4 of its decision formed part of an ongoing process following a common plan and that that conduct did not constitute isolated or sporadic events. According to the Commission, the various elements of the infringement, which pursued the same anticompetitive objective, had been developed in the context of multiple and frequent discussions between the traders of the banks concerned. By the second part of its second plea in law, Crédit agricole alleges the ‘absence of a continuous infringement’. In response to a question from the Court at the hearing, Crédit agricole stated that, in the context of points 3.16 to 3.31 of that plea, it disputed the continuous nature of its participation in the infringement, but that the interruptions to its participation and to that of the other banks should also have consequences for the continuous nature of the infringement at issue as a whole. In the third part of its second plea in law, Credit Suisse claims that the Commission has not proved that the infringement at issue continued after February 2013. The third part of the second plea in law relied on by Credit Suisse and the second part of the second plea in law relied on by Crédit agricole must be examined in turn, in so far as the second part seeks to show that the infringement at issue was not continuous. (i) The third part of Credit Suisse’s second plea in law, alleging lack of evidence of the continuing nature of the infringement at issue after February 2013 By the third part of its second plea in law, Credit Suisse claims that the Commission has not shown that the infringement at issue was continuous after February 2013, at which point there would have been a significant change in that infringement. As a preliminary point, Credit Suisse claims that, in the present case, the channel of communication between the traders in question was not clandestine and that the exchanges were perfectly recorded, so that the absence of evidence of communication for a given period constitutes effective proof that the conduct was interrupted during that period. The Commission is therefore wrong to argue that the weakness of its case is due to fragmentary and incomplete evidence. According to Credit Suisse, given the dynamics of the relevant market and the chronology of events after February 2013, the Commission should have adduced evidence relating to events that took place sufficiently close in time. After February 2013, there were numerous long gaps, including gaps varying between 49 and 82 days. Taking into account bilateral discussions between traders other than the Credit Suisse trader, there were 10 periods of inactivity of 28 days or more. In that context, Credit Suisse claims, first of all, that the fact that the conduct adopted by the banks in question pursued a common objective is not sufficient to demonstrate a continuing infringement in the presence of evidence or indications that the infringement was interrupted. Such arguments would at most support the existence of a single ‘repeated’ infringement. Similarly, the Commission simply made deductions based on a body of evidence taken as a whole, which would not demonstrate, to the requisite legal standard, a continuing infringement. Next, Credit Suisse points out that the Commission did not explain how the gaps referred to in paragraph 372 above could be linked, given that the information exchanged was rapidly becoming obsolete and that the Commission tried to prolong the effects of a few sporadic discussions identified after February 2013. Finally, according to Credit Suisse, the circumstances of the case and the operation of the market had in fact necessitated ‘positive measures’ and discussions taking place with some regularity and at relatively short intervals. The Commission is unable to refute that argument. In that regard, as a preliminary point, Credit Suisse’s argument that the channel of communication between the traders of the banks concerned was not clandestine and that the exchanges are perfectly recorded must be rejected, so that the absence of evidence of communication for a given period constitutes effective proof that the conduct was interrupted during that period. First, that argument is contradicted by certain explanations contained in the contested decision. In recital 777 of that decision, the Commission stated that, following the prohibition on the use of persistent multilateral chat rooms by Deutsche Bank, it was normal for the evidence of bilateral discussions to be sparser once the traders knew that they had to stop using multilateral chat rooms and therefore that they needed to ‘conceal their contacts from internal compliance’. Furthermore, in recitals 82 and 778 of that decision, the Commission explained that, during the investigation, it had been informed that the audio recordings of the BofA trader’s telephone conversations for the period between 29 November 2014 and 3 September 2015 had been deleted and that the only information it had, namely a directory of outgoing calls, showed that, between November 2014 and March 2015, 14 calls had taken place between the BofA trader and the Crédit agricole trader. The Commission added that only one of those calls was listed in Crédit agricole’s system, namely a call on 10 December 2014. Secondly, absence of manifestation of the infringement at issue during a given period cannot automatically be interpreted as an interruption of that infringement during that same period. Even if the manifestations of the cartel are separated by more or less long periods of time, the Commission may, in assessing the continuous nature of an infringement extending over several years, consider that the infringement has not been interrupted, first, where the various actions forming part of the infringement pursue a single purpose and are capable of forming part of a single infringement and, secondly, where there is no indication or evidence that the infringement did not continue during those periods (see, to that effect, judgments of 18 March 2021, Pometon v Commission , C‑440/19 P, EU:C:2021:214 , paragraphs 112 and 114, and of 17 May 2013, Trelleborg Industrie and Trelleborg v Commission , T‑147/09 and T‑148/09, EU:T:2013:259, paragraphs 61 and 63). However, the principle of legal certainty requires the Commission at least to rely on evidence of facts sufficiently proximate in time for it to be reasonable to accept that that infringement continued uninterruptedly between two specific dates (see, to that effect, judgment of 29 September 2021, Nichicon Corporation v Commission , T‑342/18, EU:T:2021:635 , paragraph 365 (not published) and the case-law cited). While the period separating two manifestations of infringing conduct is a relevant criterion for establishing the continuous nature of an infringement, the question whether or not that period is long enough to constitute an interruption of the infringement cannot be examined in the abstract. On the contrary, it needs to be assessed in the context of the functioning of the cartel in question (see judgment of 29 September 2021, Nichicon Corporation v Commission , T‑342/18, EU:T:2021:635 , paragraph 366 (not published) and the case-law cited). In the present case, in the first place, it is clear from paragraphs 317 to 362 above that, throughout the infringement period, the participants’ conduct formed part of an overall plan pursuing a single anticompetitive objective. It follows that the Commission could consider that the infringement at issue had been continuous throughout the entire infringement period set out in the contested decision, unless it were to be found that the infringement had been interrupted (see paragraphs 379 to 381 above). In the second place, the anticompetitive objective pursued was to maximise the revenues of the banks concerned while limiting their losses, which could arise as a result of uncertainty as to how other traders would behave. The conduct implemented concerned various SSA bonds and did not consist solely of coordination on prices or exchanges of information on prices. That conduct concerned, more broadly, the trading activities of the banks concerned. As is clear from paragraphs 323 and 357 above, the pursuit of the single anticompetitive objective found in the contested decision did not, contrary to what Credit Suisse suggests, depend on the organisation of daily discussions. The purpose of the discussions between the traders concerned was to identify and exploit opportunities to coordinate or to exchange commercially sensitive information. Furthermore, and in the light of the foregoing, very frequent participation in collusive contacts was not necessary for the operation of the infringement at issue, given that the SSA bond negotiations were independent. The possible success of a given collusive contact did not depend on participation in a series of contacts relating to other negotiations. Finally, the discussions that took place after February 2013 involved the same traders who were behind the setting up of persistent multilateral chat rooms in 2010. Those traders met frequently at social events and formed a close-knit group that was well known among traders in the City of London. In the third place, the Commission has relied on evidence which relates to events which took place sufficiently close together in time, and the gaps relied on by Credit Suisse between two manifestations of the infringement or between two discussions which the Commission is able to prove do not allow it to be considered that the infringement was interrupted in the period after February 2013. In that regard, first, it should be pointed out that Credit Suisse does not allege that the infringement was interrupted before February 2013, although several gaps of more than one month between certain discussions can be observed before that date and, more specifically, in 2012. Secondly, in recital 534 of the contested decision, the Commission referred to several documents from which it emerged that, during 2014 and 2015, the traders perceived the SSA bond market as becoming increasingly difficult and less lucrative (‘no money in SSA anymore’, ‘no flows in SSA’, ‘no money in SSA’, ‘SSA is dead’). That phenomenon explained why the traders in question were increasingly trading other financial instruments to compensate for the reduction in activity on that market. That reduction in activity, which was beyond the control of the participants in the single infringement, contributed to a reduction in opportunities to coordinate or exchange commercially sensitive information. Thirdly, an examination of the discussions which preceded the gaps invoked by Credit Suisse and of the discussions which followed those gaps shows that there is nothing in those discussions to demonstrate that the traders of the banks concerned initially showed a willingness to put an end to the single infringement and then a willingness to carry it out again. By contrast, the tone used by the traders was particularly free and the discussions were spontaneous. In particular, the anticompetitive discussions that followed the gaps invoked by Credit Suisse did not contain any introductory remarks that would demonstrate the need to re-establish contact, or even trust, between the participants that had previously been broken. Those discussions thus establish a persistent and constant willingness on the part of the participating traders to engage in conduct that contributed to an overall plan pursuing a single anticompetitive objective, as defined by the Commission in the contested decision. More specifically, with regard to the 64-day gap between, first, the discussion of 21 March 2013 between the Crédit agricole trader and the Deutsche Bank trader and, secondly, the discussion of 24 May 2013 between the Crédit agricole trader and the Credit Suisse trader, it should be noted that, although the obligations concerned by those discussions were different, the content of the discussions was the same. The traders in question exchanged commercially sensitive information about trading activities that were under way or had just taken place, and in particular about their prices and clients (see paragraphs 217 to 227 and 228 to 232 above). Moreover, the comments made during those discussions are particularly free and the traders’ conduct is natural. There is nothing in those discussions to suggest that the traders’ intention to carry out the infringement ceased between those two manifestations of the infringement. As regards the 60-day gap between the two discussions of 11 October 2013 and 10 December 2013 between the Credit Suisse trader and the Deutsche Bank trader, it should be noted that the two discussions related to different bonds, but that the content of the discussions was similar. During the first of those discussions, the Credit Suisse trader withdrew a sell order from a broker’s platform (‘killed it’) after the Deutsche Bank trader suggested that withdrawal because of his long position (‘i am long bro … if ok to scarp it that would be better’). In the second of those discussions, it was the Deutsche Bank trader who suggested withdrawing his bid (‘ah sorry bro … will kill it’) after learning of the Credit Suisse trader’s short position (‘i’m small short’). The fact that the Deutsche Bank trader finally decided not to withdraw his bid after hearing the opinion of the Credit Suisse trader does not call into question the assessment according to which the said discussions show continuity in their content and a form of reciprocity over time with regard to the withdrawal and a proposal to withdraw a bid that could harm another participant in the cartel. Furthermore, those discussions do not mention any element that would suggest that the traders’ willingness to implement the infringement had ceased between those two manifestations of the infringement. The first of those discussions, namely that of 11 October 2013, at which traders from the three banks involved were present, also showed a desire to schedule a meeting of a social nature during a business trip to New York (United States of America) the following week. As regards the 55-day gap between the two discussions of 10 December 2013 and 3 February 2014 between the Credit Suisse trader and the Deutsche Bank trader, it should be noted that those two discussions related to different bonds, but that the content of those discussions was similar. During the first of those discussions, the Deutsche Bank trader suggested withdrawing his bid after learning of the Credit Suisse trader’s short position (see paragraph 393 above). During the second of those discussions, the Deutsche Bank trader refrained from hitting a bid submitted by Credit Suisse on a broker’s platform and therefore from selling him a bond after an exchange between the two traders. The two discussions were direct and spontaneous, and there was nothing to suggest that the two traders had ceased to be willing to contribute to the infringement at issue between those two manifestations of the infringement. As regards the 82-day gap between, first, a discussion on 1 May 2014 between the Crédit agricole trader and the Credit Suisse trader and, secondly, a discussion on 22 July 2014 between the Credit Suisse trader and the BofA trader, it should be pointed out that, during the first of those discussions, the Crédit agricole trader and the Credit Suisse trader discussed the price of a particular bond (KFW 5 1/8 03/16). The Crédit agricole trader asked ‘where u ma[rk]ing that sh1t[?]’. The Credit Suisse trader explained ‘got them marked +2 on the bid (2/-2)’, to which the Crédit agricole trader replied ‘perfect’. That exchange enabled the Crédit agricole trader to check his prices with the Credit Suisse trader. During the second of those discussions, the BofA trader, who had previously taken part in discussions with the other traders when he worked for Deutsche Bank, suggested to the Credit Suisse trader that he withdraw his bid for another bond. The BofA trader also mentioned his trading position (‘sorry I didn’t realise kbn 19 in tullets was you … i went better bid … can kill it if you [want] … we are short 10mm …’). During that 82-day gap, that former Deutsche Bank trader had been on leave since 1 April, in anticipation of leaving that bank in June 2014 and joining BofA. That event did not have the effect of interrupting the infringement, since on 1 May 2014 the Crédit agricole trader and the Credit Suisse trader discussed the price of a bond (see paragraph 395 above). By contrast, the Deutsche Bank trader’s leave, prior to his change of employer, contributed to a temporary reduction in opportunities for coordination or exchanges of commercially sensitive information between the three traders principally involved. The former Deutsche Bank trader then took up his position at BofA on 16 July 2014. The fact that that trader, six days later, namely on 22 July 2014, held a bilateral discussion with the Credit Suisse trader in order to offer to withdraw an bid (‘can kill it if you [want]’) and to set out his negotiating position (‘we are short 10 mm’) demonstrates an uninterrupted willingness to engage in conduct falling within the overall plan pursuing a single anticompetitive objective defined by the Commission. That continued willingness is also illustrated by the direct tone used by the two participants and the spontaneity of the exchange. As regards the gap of 67 days relied on by Credit Suisse between an alleged discussion on 24 August 2014 and a discussion on 29 October 2014, it should be pointed out that the argument which Credit Suisse is trying to draw from it is unfounded in so far as no bilateral discussion was identified on 24 August 2014. The discussion referred to by Credit Suisse actually dates from 24 September 2014 and the gap relied on is therefore 37 days. In any event, those discussions do not show any interruption of the infringement and, on the contrary, show that, on those two occasions, the Crédit agricole trader and the BofA trader continued to exchange information on their pricing strategy and thus demonstrated their persistent and constant willingness to fall within the single infringement retained by the Commission in the contested decision. As regards the 70-day gap between, first, a discussion on 29 October 2014 between the Crédit agricole trader and the BofA trader and, secondly, a discussion on 7 January 2015 between the Crédit agricole trader and the Credit Suisse trader, there is nothing in the content of the first of those discussions to suggest a willingness on the part of the participants to interrupt the infringement at issue, while there is nothing in the content of the second of those discussions to suggest a desire to re-implement an infringement that had been interrupted. Those two discussions show an exchange of information on prices. Those discussions therefore illustrate a common and persistent desire to continue to implement the overall plan defined by the Commission in the contested decision. Finally, with regard to the 45-day gap between, first, a discussion on 27 January 2015 between the BofA trader and the Crédit agricole trader and, secondly, a discussion on 12 March 2015 between the Crédit agricole trader and the Credit Suisse trader, it is important to note that the content of the discussion on 27 January 2015 shows an exchange about the prices of a specific bond. During that discussion, the BofA trader asked his interlocutor how he valued the said bond and the Crédit agricole trader replied. There is no indication in that discussion that the BofA and Crédit agricole traders considered ending their respective conduct in the future. In that regard, the fact that the discussion of 27 January 2015 was the last anticompetitive discussion held against BofA and that it marked the end of that bank’s participation in the infringement at issue cannot be interpreted as an interruption of the single infringement carried out by the other two participating traders, namely the Crédit agricole trader and the Credit Suisse trader. Thus, the content of the discussion of 12 March 2015 between the Crédit agricole trader and the Credit Suisse trader bears witness to an exchange of very precise and up-to-date information on an offer made to a client, but also to coordination of the traders’ prices and trading activity (see paragraphs 257 to 267 and 296 to 300 above). That exchange of information and that coordination are not preceded by any introductory statement that would suggest that the single infringement had previously been interrupted. On the contrary, the exchange of information and the coordination took place a few seconds after the participants entered the non-persistent chat room. The assessments set out in paragraph 391 above also apply to the other gaps identified by Credit Suisse between the other discussions that took place after February 2013. It follows from the foregoing that the context of the operation of the cartel found, which lasted for approximately five years and two months, between 19 January 2010 and 24 March 2015, supports the conclusion that the single infringement was continuous despite the gaps relied on by Credit Suisse. On the one hand, the traders of the banks concerned pursued the anticompetitive objective found by the Commission in the contested decision by freely exchanging information about their ongoing trading activities. Those exchanges took the form of a continuous, real-time description of events as they occurred, which enabled them to identify and exploit opportunities to coordinate and exchange commercially sensitive information when the opportunity arose and, in particular, when those participating traders showed a common interest in one or more specific SSA bonds. Those traders thus sought to achieve the said anticompetitive objective in the context of anticompetitive exchanges relating to given negotiations which were independent of each other, or even in the context of anticompetitive exchanges relating to information relevant to given periods. It was therefore not necessary to have ongoing or daily discussions in order to pursue the anticompetitive objective in question. On the other hand, several events from February 2013 could have affected the willingness of all participating traders to adhere to the overall plan pursuing a single anticompetitive objective, namely, first, the prohibition on Deutsche Bank traders from using persistent multilateral chat rooms in a context of increased supervision in the investment banking sector, secondly, the garden leave of several months of the Deutsche Bank trader prior to his taking up his post at BofA, or thirdly, the decline in activity on the SSA bond market. However, despite those events, the discussions identified by the Commission do not reveal any questioning of the cartel or, in particular, any questioning on the part of the participating traders as to their willingness to continue their anticompetitive conduct. The traders of the banks concerned continued, albeit less frequently but repeatedly, their discussions of an anticompetitive nature relating to specific negotiations which were independent of each other. The reaction of those traders can be explained by the fact that the anticompetitive objective they were pursuing was based on their common understanding that they were helping each other in order to derive mutual benefit over time. Thus, on that basis, a trader could refrain from acting on a specific transaction or share a transaction with another trader and thus forego an immediate benefit. A trader could also exchange information with other traders, namely his or her competitors, and thus forego taking immediate advantage of that information on his or her own. In effect, a trader would refrain from acting, share a transaction or exchange commercially sensitive information on the grounds that he or she expected other traders to behave in the same way in the future. In the context of the operation of the cartel at issue, the conduct found by the Commission and, more generally, the overall plan in which that conduct formed part were therefore based on mutual expectations which were fed by each anticompetitive discussion and the purpose of which was to derive a lasting advantage vis-à-vis clients and competing traders and, ultimately, to maximise the income of the participating traders while limiting their losses (see, to that effect, recitals 643, 744, 746, 781 and 810 of the contested decision). In those circumstances, the shortness of the time limits within which the information exchanged became obsolete, invoked by Credit Suisse, did not require the Commission to explain how the gaps found between two manifestations of the infringement had been remedied. Moreover, that brevity is not such as to call into question the Commission’s conclusion that the single infringement at issue was of a continuing nature. In the light of the foregoing, Credit Suisse’s arguments alleging, in essence, that the Commission did not adduce evidence of facts sufficiently close in time to conclude that there was a continuous infringement cannot succeed. That conclusion is not called into question by the judgment of 10 November 2017, Icap and Others v Commission (T‑180/15, EU:T:2017:795 ), relied on by Credit Suisse. First of all, it should be pointed out that, in the case which gave rise to the judgment of 10 November 2017, Icap and Others v Commission (T‑180/15, EU:T:2017:795 ), the Court assessed the continuous nature of the participation of a particular undertaking and not the continuous nature of the infringement as a whole. Next, the object of the cartel and its operation were different from those of the infringement at issue. In the judgment of 10 November 2017, Icap and Others v Commission (T‑180/15, EU:T:2017:795 ), the Court took into account, as part of the context of the operation of the infringements at issue, the daily nature of the fixing of JPY LIBOR rates, namely a set of reference interest rates which was used for numerous financial products denominated in Japanese yen. It was in that context that the Court considered that the Commission should highlight positive measures adopted on a basis that was, if not daily, at least sufficiently limited in time. By comparison, in the present case, having regard to the context of the operation of the cartel and, more specifically, having regard to the fact that that cartel was based on recurring collaboration on specific SSA bonds, it was not necessary for the Commission to highlight ‘positive measures’ adopted by the cartel participants in as frequent a manner as in the case which gave rise to the judgment of 10 November 2017, Icap and Others v Commission (T‑180/15, EU:T:2017:795 ). In any event, it should be noted that the position of the participants in the infringement at issue cannot be compared to that of Icap in the case which gave rise to the judgment of 10 November 2017, Icap and Others v Commission (T‑180/15, EU:T:2017:795 ). Icap had acted as a ‘facilitator’ of a cartel between banks and its economic activity was not the same as that of those banks. In particular, the Court noted that the Commission had relied on Icap’s implementation of infringements decided each time between two banks and that the operation of the cartel made it more difficult for the Commission to show that Icap should reasonably have deduced that the requests made to it by one bank were part of a collusion with another bank. In the present case, the Commission found that the banks involved had participated directly in at least some aspects of the single infringement found. Similarly, given the differences which characterise the context of the operation of the infringement at issue and the context of the operation of the cartel which was at issue in the case which gave rise to the judgment of 7 November 2019, Campine and Campine Recycling v Commission (T‑240/17, not published, EU:T:2019:778 ), in particular as regards the product concerned and the importance of the price factor, the assessments made by the Court in that judgment cannot be transposed to the present case. Finally, in so far as Credit Suisse relies on the judgment of 14 January 2021, Kilpailu- ja kuluttajavirasto (C‑450/19, EU:C:2021:10 ), and on the Opinion of Advocate General Pitruzzella in Kilpailu- ja kuluttajavirasto (C‑450/19, EU:C:2020:698), it should be pointed out that that judgment sought to determine the end date of a single and continuous infringement in order to enable the national court hearing the main proceedings to determine the starting point of the limitation period. The case which gave rise to that judgment did not therefore concern the identification of any provisional interruption of a single infringement and the assessment of the continuing nature of that infringement. Furthermore, it is not apparent from the case-law referred to in paragraphs 379 to 381 above, which refers in particular to the context of the operation of the cartel, that the protected legal interest should, as Credit Suisse submits on the basis of the judgment of 14 January 2021, Kilpailu- ja kuluttajavirasto (C‑450/19, EU:C:2021:10 ), be taken into account in assessing the existence of a temporary interruption of a single infringement. The third part of the second plea in law put forward by Credit Suisse is therefore rejected. (ii) The second part of Crédit agricole’s second plea in law, in so far as it seeks to demonstrate that the infringement was not continuous Crédit agricole argues that the Commission did not adequately take into account the change in the nature and frequency of discussions towards the end of 2013 and the beginning of 2014. The Commission allegedly tried to explain this by referring to the declining profitability of the SSA bond market and the replacement of persistent multilateral chat rooms by bilateral discussions, which would explain why there was less evidence available. However, according to Crédit agricole, that replacement should probably have given rise to more evidence, given that a single multilateral chat room was replaced by several bilateral chat rooms. In addition, the banks concerned, other than the immunity applicant in the present case, also provided contemporaneous documentary evidence, so the drop in the volume of evidence would simply reflect a net reduction in the number of discussions from October 2013. That reduction was due in particular to a long period of absence on the part of the Deutsche Bank trader, after which there was a single instance of ‘coordination’ on 12 March 2015, which incidentally constituted a misinterpretation by the Commission of an exchange involving Credit Suisse. As a preliminary point, for the reasons set out in paragraphs 376 to 381 above, Crédit agricole’s argument that the gaps between the anticompetitive discussions constitute periods during which the conduct at issue did not take place should be rejected. First, as regards the continuing nature of the infringement, it should be noted that, with the possible exception of the arguments rejected in paragraphs 363 and 364 above, Crédit agricole has not contested the Commission’s findings relating to the existence of an overall plan pursuing a single anticompetitive objective. In addition, Crédit agricole admittedly noted that anticompetitive discussions were even less frequent after October 2013, in particular because Deutsche Bank had banned the use of persistent multilateral chat rooms. However, Crédit agricole does not argue that that circumstance would have justified calling into question the overall plan in question or the objective elements supporting the existence of that plan. In any event, the mechanism and pattern common to the anticompetitive discussions covered by the infringement at issue were not called into question by that prohibition and by the switch to a network of bilateral discussions (see paragraphs 357 to 360 above). More generally, the Commission did not commit any errors with regard to the existence of an overall plan pursuing a single anticompetitive objective until March 2015 and the identification of five categories of conduct is not likely to call into question the existence of that overall plan (see paragraphs 70 to 89 and 364 above). It follows that the Commission could consider that the infringement at issue had been continuous throughout the entire infringement period set out in the contested decision, unless the infringement was found to have been interrupted (see paragraphs 379 to 381 above). Secondly, with regard to the other elements discussed between the parties, it should be noted that, although Crédit agricole referred to the Commission’s reason based on the decline in the profitability of the SSA bond market, set out in particular in recital 534 of the contested decision, in order to contest its relevance, that bank did not contest the reality of that circumstance. Moreover, although Crédit agricole refers to the reduced participation of the Deutsche Bank trader during a certain period, it does not dispute that this was the result of his ‘garden leave’, namely the period during which, before joining BofA, the trader was still employed by Deutsche Bank but was not working. Nor has Crédit agricole disputed that the trader resumed taking part in anticompetitive discussions shortly after joining BofA (see paragraphs 396 and 397 above and recitals 534, 557 and 558 of the contested decision). In the light of the foregoing and in the absence of evidence or indications to the contrary, although it is possible that the infringement at issue was less serious, the evidence put forward by Crédit agricole is not such as to affect the continuing nature of that infringement, and therefore does not allow the conclusion to be drawn that there was an interruption in the commission of the infringement at issue. That is not called into question by the considerations in the contested decision relating to ‘constant communications’ (recitals 95, 96 and 761). On the one hand, it is not disputed that the traders of the banks concerned were relatively close, on a personal level, and that they conversed frequently, irrespective of whether or not those discussions were anticompetitive. On the other hand, Crédit agricole does not dispute that, at least during certain periods prior to October 2013, the anticompetitive discussions had been frequent. Thirdly, the other arguments put forward by Crédit agricole are not such as to call into question the continuing nature of the single infringement found by the Commission. First, the rest of Crédit agricole’s argument is based on its individual situation. More specifically, the rest of that bank’s argument is based on the length of the periods between the discussions in which it took part and which the Commission considered to be anticompetitive or which Crédit agricole acknowledges, for the purposes of that part of its argument, to be anticompetitive. Crédit agricole’s argument therefore largely ignores the discussions which, according to the Commission, form part of the single infringement and which involved traders from the other two banks concerned. In particular, the gaps invoked by Crédit agricole between two manifestations of the infringement do not take account of the conduct of the other participants. Secondly, Crédit agricole has not shown that the absence of its participation in the single infringement during certain periods is likely to call into question the continuing nature of that infringement as a whole. Accordingly, Crédit agricole’s argument, in so far as it is based on the length of the periods separating the discussions in which it participated, is not capable of calling into question the Commission’s conclusion that the single infringement identified in the contested decision was of a continuous nature. By contrast, that argument is capable of demonstrating an interruption in Crédit agricole’s ‘participation’ in that single infringement and it is therefore in that context that that argument will be examined subsequently. The second part of Crédit agricole’s second plea in law is therefore rejected in so far as it seeks to show that the single infringement found by the Commission was not continuous. (d) The participation of Crédit agricole and Credit Suisse in the single and continuous infringement found It should be borne in mind that the finding of the existence of a single and continuous infringement is distinct from the question whether liability for that infringement as a whole is attributable to an undertaking. Moreover, whether liability for the single and continuous infringement as a whole is imputable to an undertaking must be assessed in the light of two factors, namely, first, the intentional contribution of that undertaking to the common objectives pursued by all the participants and, secondly, its knowledge of the infringing conduct planned or implemented by other undertakings in pursuit of the same objectives or the fact that it could reasonably have foreseen it and had been prepared to take the risk (see paragraphs 148 to 153 above). In recitals 780 to 787 of the contested decision, the Commission set out the reasons why it considered that the banks concerned had intentionally contributed to the overall plan. In particular, it explained that it was clear from the evidence presented in Section 4.2 of the contested decision, devoted to the chronological presentation of a selection of more than 200 discussions which took place between 19 January 2010 and 12 March 2015, that each of the banks concerned had intentionally contributed to the implementation of the overall plan pursuing a single anticompetitive objective through the actions of their employees during one or more given periods. In recitals 807 to 826 of the contested decision, the Commission set out the reasons why it considered that each of the banks concerned had knowledge of the anticompetitive conduct of the others. As regards, in particular, the period after February 2013, namely after Deutsche Bank had prohibited its traders from using persistent multilateral chat rooms, the Commission noted, in essence, that the traders of the banks concerned had continued to keep each other informed by using a network of bilateral chat rooms, whereby, in the context of given bilateral discussions, two traders of the banks concerned who were present communicated the information obtained from the third trader who was absent. It added that the information had concerned the same types of items as those exchanged during the discussions in the multilateral chat rooms. The first part of the second plea put forward by Crédit agricole formally alleges that it did not contribute to an overall plan. For its part, Credit Suisse claims, in the second part of its second plea in law, that the Commission has not established that it participated directly in the bilateral discussions between the other banks after February 2013, or that it was aware of them or could reasonably have foreseen them. (1) The first part of the second plea in law put forward by Crédit agricole, formally alleging that it did not contribute to an overall plan The first part of the second plea in law put forward by Crédit agricole formally alleges that it did not contribute to an overall plan. However, in the context of that plea, it also alleges ignorance of certain conduct engaged in by the other banks concerned. Crédit agricole’s arguments should be examined by distinguishing between, on the one hand, its intentional contribution to an overall plan and, on the other hand, its knowledge of all the conduct implemented or envisaged by the other banks concerned. (i) Crédit agricole’s intentional contribution to an overall plan First, Crédit agricole disputes that it intended to contribute to an overall plan pursuing a common objective, in that it did not seek to maximise its revenues, but rather to establish a reputation on the secondary market for SSA bonds. Secondly, Crédit agricole claims that, in so far as it did not participate in certain categories of conduct constituting the single and continuous infringement at issue, the Commission would have been required to show that it was aware or should have been aware that its participation in the other categories formed part of a wider plan. According to Crédit agricole, the discussions held by the Commission with regard to it in fact demonstrate legitimate attempts to find liquidity, seek prices and carry out transactions. On that point, Crédit agricole refers to the arguments it put forward in its first plea in law. Crédit agricole adds that, even if it accepts the description of ‘coordination discussions’, the Commission has not shown that it participated in the first category of coordination, namely the coordination on prices quoted to specific clients. In that regard, in the first place, it should be noted that Crédit agricole did not explain what its objective of establishing its ‘reputation’ on the secondary market in SSA bonds consisted of, nor in what way that objective was incompatible with the objective of the other traders of the banks concerned of reducing uncertainties as to the conduct which they would respectively adopt in given negotiations, by means of coordination and exchanges of information (recital 757 of the contested decision). Furthermore, it must be noted that Crédit agricole did not contest the findings made in recitals 780 and 781 of the contested decision relating to the intention of the banks concerned to contribute to the overall plan, in particular the findings relating to the active participation of those banks’ traders in the multilateral and bilateral chat rooms. Nor did Crédit agricole dispute the fact that its trader joined the multilateral chat room shortly after joining that bank, following the termination of his duties at BofA (recital 784 of, and footnote 891 to, the contested decision). In the second place, Crédit agricole’s argument is first of all inoperative in that it seeks to challenge its intentional contribution on the basis of the erroneous premiss that the Commission found five separate infringements corresponding to the five categories of conduct identified by the Commission in the contested decision (see paragraphs 87 and 89 above). Next, the Commission was entitled to consider that Crédit agricole was aware of the discussions of 31 January and 11 October 2013, which took place in its presence and in which it therefore participated (see paragraphs 133 and 134 and 196 to 200 above). By contrast, the Commission could not validly hold the discussion of 10 January 2013 against that bank (see paragraph 145 above). Finally, it is clear, first, from the discussions that Crédit agricole is not entitled to contest and, secondly, from the discussions that that bank is entitled to contest and, as examined in paragraphs 188 to 268 above, that its trader contributed to around 30 anticompetitive discussions between 11 January 2013 and 24 March 2015 and, in particular, to the discussion of 12 March 2015 which establishes the existence of category 1 conduct, namely coordination on prices quoted to specific counterparties. In the third place, the contribution of the Crédit agricole trader to the anticompetitive objective pursued by all the participants was intentional. The anticompetitive conduct in which the Crédit agricole trader participated took the form of coordination of prices and trading activity. In that context, he provided assistance to competing traders or received assistance from the traders concerned in the common interest of all the participants. Furthermore, the Crédit agricole trader took part in anticompetitive exchanges of information. In that regard, the Crédit agricole trader was necessarily aware of the fact that, in the context of normal commercial transactions, the disclosure to competing traders of information about his prices, clients or ongoing negotiations could have weakened his position. However, in the present case, the trader had the assurance that the information he communicated to other traders would not be used to his detriment and that the exchanges of information in which he participated would give him an advantage and enable him to coordinate his activities with the other traders. Consequently, Crédit agricole has not shown that the Commission erred when it considered that that bank’s trader had intentionally contributed to the achievement of the single anticompetitive objective pursued by all the participants and that the discussions covered all the categories of conduct identified by the Commission. (ii) Crédit agricole’s knowledge of all the other conduct implemented or envisaged by the other banks concerned Crédit agricole maintains that it could not have been aware of the conduct falling within category 1. In order to demonstrate that it had knowledge of conduct falling within that category, the Commission merely made vague statements to the effect that the traders of the banks concerned knew each other and must therefore have had knowledge of such conduct. In addition, certain scattered passages of Crédit agricole’s pleadings suggest that it disputes, more generally, its knowledge of the conduct adopted by the other banks concerned and the fact that it could reasonably have foreseen it, particularly after Deutsche Bank traders ceased using persistent multilateral chat rooms in February 2013. In that regard, in the first place, as regards Crédit agricole’s claim that it was not aware of the existence of conduct falling within category 1 on the ground that such conduct was distinguishable from the other conduct at issue, it is sufficient to point out that the Commission did not find five separate infringements corresponding to each of the five categories of conduct identified in the contested decision (see paragraphs 70 to 89 above). The Commission did not therefore have to establish Crédit agricole’s precise knowledge of each of the five categories of conduct, taken separately. In any event, it is clear from paragraphs 257 to 268 above that the Crédit agricole trader participated directly in the discussion of 12 March 2015, which establishes the existence of category 1 conduct, namely coordination on prices quoted to specific counterparties. In the second place, assuming that, by its arguments, Crédit agricole disputed that it was aware of or could reasonably have foreseen all of the conduct of the other banks concerned, particularly as regards category 1, the following findings should be made. First, in the course of his duties at Crédit agricole, the trader at that bank participated directly in anticompetitive discussions which took place between two or three traders within persistent multilateral chat rooms on 11, 18, 23 and 31 January 2013 and on 15 February 2013. Those discussions gave rise to conduct which, according to the Commission, fell within categories 2, 3, 4 and 5. Crédit agricole’s criticisms of the Commission’s assessments of the discussions of 18 and 31 January 2013 and the discussion of 15 February 2013 were rejected in paragraphs 188 to 208 above. Furthermore, the Commission’s assessments that Crédit agricole participated in several other anticompetitive discussions must be considered definitive in so far as those discussions were not duly contested by that bank. Secondly, Crédit agricole does not dispute that it was aware that Deutsche Bank traders had stopped using the persistent multilateral chat rooms as a result of a prohibition imposed on them by their employer in February 2013. Thirdly, Crédit agricole participated directly in bilateral anticompetitive exchanges after February 2013 with the other banks concerned. Those exchanges covered all of the conduct identified by the Commission and, in particular, category 1 conduct. On the one hand, the Crédit agricole trader took part in anticompetitive bilateral discussions with the Deutsche Bank trader and then the BofA trader on 21 March, 3 June, 2, 10, 12 and 25 July and 4 September 2013, on 6 August, 24 September and 29 October 2014, and on 7 and 27 January 2015. On the other hand, the Crédit agricole trader participated in anticompetitive bilateral discussions with the Credit Suisse trader on 19 March, 24 May and 9 and 14 August 2013, on 4, 7 and 14 March, 1 May and 4 August 2014, and on 12 and 24 March 2015. Until August 2013, the Crédit agricole and Credit Suisse traders continued to use the persistent multilateral chat room that had been used before and during February 2013 by all the banks concerned. Finally, on 24 March 2015, the Crédit agricole trader received an axe sheet from the Credit Suisse trader. Fourthly, on 11 October 2013, the Deutsche Bank trader invited the Credit agricole trader to a chat room in which the Deutsche Bank trader and the Credit Suisse trader were taking part. During an exchange in which the Crédit agricole trader was present, the Credit Suisse trader withdrew a sale order filed with a broker after the Deutsche Bank trader had suggested that he do so (‘i am long bro … if ok to scarp it that would be better’). As is clear from paragraph 134 above, Crédit agricole has not shown that its trader did not become aware of the anticompetitive discussions at issue that took place on 11 October 2013 between 09:18:25 and 09:19:15 on the CHAT-fs:5257AB6D 02E00121 chat room, to which that bank’s trader had been connected since 08:44:58 on the same day. Fifthly, other discussions mentioned by the Commission in the contested decision show that the Crédit agricole trader was aware of the existence of discussions between the Deutsche Bank trader and the Credit Suisse trader and of the fact that those discussions related in particular to their respective trading activities. During a discussion on 10 July 2013, the Deutsche Bank trader informed the Crédit agricole trader of the situation of the Credit Suisse trader (‘[Credit Suisse’s trader has] also been hit’) (recital 501 of the contested decision). Similarly, during a discussion on 22 October 2013, referred to in recital 823 of the contested decision, the Crédit agricole trader asked the Deutsche Bank trader whether he still had any ‘ifc 18’ bonds (‘u got any ifc 18 left’). The Deutsche Bank trader replied in the negative and said that he had sold the last piece to the Credit Suisse trader who might still have them (‘nope … all sold … sold last piece to [Credit Suisse’s trader] … he might still have them’). The Crédit agricole trader replied that he would ask him (‘ok will ask him’). Thus, because of its participation in the persistent multilateral chat rooms before 25 February 2013, Crédit agricole knew that, before that date, Deutsche Bank and Credit Suisse traders were colluding. Moreover, given that, after that date, the Crédit agricole trader took part in anticompetitive bilateral discussions with each of those two traders and that the Commission had evidence to show the existence, after 25 February 2013, of a network of bilateral discussions in which the three traders continued to discuss their trading activities, it must be considered that the Crédit agricole trader could, at the very least, reasonably foresee that the other two traders were conducting bilateral discussions which were anticompetitive in nature. The finding that Crédit agricole could, at the very least, reasonably foresee all of the conduct of the other banks and, in particular, the coordination on prices quoted to specific counterparties is also corroborated by other elements. First, it should be borne in mind that an employee performs his or her duties for and under the direction of the undertaking for which he or she works and is thus regarded as forming part of the economic unit constituted by that undertaking (see judgment of 21 July 2016, VM Remonts and Others , C‑542/14, EU:C:2016:578 , paragraph 23 and the case-law cited). Knowledge acquired by an employee prior to joining a new undertaking and which the employee in fact makes available to the new employer may therefore be regarded as knowledge shared by his or her new employer. Moreover, according to settled case-law, the Commission may rely on contacts prior to or subsequent to the period of the infringement in order to construct an overall picture and to show the preparatory stages of the cartel as well as to corroborate the interpretation of certain evidence (see, to that effect, judgments of 8 July 2008, Lafarge v Commission , T‑54/03, not published, EU:T:2008:255, paragraph 428; of 2 February 2012, Denki Kagaku Kogyo and Denka Chemicals v Commission , T‑83/08, not published, EU:T:2012:48, paragraph 188; and of 27 June 2012, Coats Holdings v Commission , T‑439/07, EU:T:2012:320, paragraph 60). The Commission is therefore entitled to use the knowledge acquired by an employee prior to his or her arrival in the service of a new company and which he or she in fact makes available to that new employer when that knowledge corroborates other elements available to the Commission. In the present case, before taking up his post at Crédit agricole, the trader at that bank had, in his capacity as a trader at BofA, participated directly in the conduct that took place within persistent multilateral chat rooms and during bilateral discussions. In addition, the conduct adopted by the Crédit agricole trader in his previous capacity fell within all the categories identified by the Commission, and in particular category 1 relating to the coordination on prices quoted to specific counterparties. The knowledge acquired by the Crédit agricole trader in the course of his previous duties is therefore likely to corroborate the other elements retained by the Commission in order to conclude that Crédit agricole could, at the very least, reasonably foresee all the conduct of the other banks. Secondly, the Commission found that the Crédit agricole trader was one of a small circle of traders particularly involved in the infringement at issue. Those traders were associated professionally and personally and had been targeted as a group (see recitals 425, 807, 809 and 810 of the contested decision). Thirdly, as the Commission stated in recital 809 of, and footnote 931 to, the contested decision, it is apparent from two parallel discussions of 23 May 2013, produced by the Commission pursuant to an investigative measure and taking place between, on the one hand, the Crédit agricole trader and the Credit Suisse trader and, on the other hand, the Credit Suisse trader and the Deutsche Bank trader, that the traders of the banks concerned were perceived as a ‘gang’, including after February 2013. During his discussion with the Crédit agricole trader, the Credit Suisse trader described meeting someone who asked him if he was part of the ‘gang’, to which he ironically replied that he said nothing to the Deutsche Bank and Crédit agricole traders (‘met that … girl last night … told her we should chat/help each other out etc. She turned around and said “but your part of the ‘gang’, can I trust you?” she says, everyone talks about [our tight-knit group] … lol … I told her, I don’t tell [Deutsche Bank’s trader] or [Crédit agricole’s trader] anything … trust me!!’). The Crédit agricole trader laughed in response. A few minutes later, as part of his side discussion with the Deutsche Bank trader, the Credit Suisse trader explained that he had just spoken with the Crédit agricole trader and reproduced the content of the message he had sent to the latter about his meeting the previous day. The Deutsche Bank trader replied with a laugh. Thus, the Commission rightly considered that Crédit agricole could, at the very least, reasonably foresee all the conduct implemented or envisaged by the other banks concerned and, in particular, the fact that that conduct concerned all the categories of conduct described by the Commission in the contested decision. In particular, the Commission rightly considered that Crédit agricole could, at the very least, reasonably have foreseen the bilateral discussions of 25 September 2013 and 5 February 2014 involving banks other than Crédit agricole, namely Credit Suisse and Deutsche Bank, which fall within category 1 (see recitals 528, 529, 539 and 540 of the contested decision). Consequently, Crédit agricole is not entitled to argue that it could not reasonably have foreseen all of the conduct of the other traders. The first part of the second plea in law put forward by Crédit agricole is therefore rejected. (2) The second part of Credit Suisse’s second plea in law, alleging errors as to its knowledge of the bilateral discussions between traders of other banks or as to the fact that it could reasonably have foreseen them Credit Suisse contests the Commission’s finding that it was liable on the basis, first, of the direct participation of that bank in all the conduct making up the infringement at issue and, secondly, of the fact that that bank had been aware of the bilateral discussions involving traders from other banks or that it could reasonably have foreseen them and had been prepared to take the risk. Those two premisses are erroneous. In the first place, Credit Suisse claims that, even if it were to be shown that it had participated in each ‘type’ of conduct, that would not be sufficient to establish its participation in each anticompetitive discussion, nor to hold it liable for discussions in which it was not directly involved. In the second place, given that the bilateral discussions after February 2013 could not, according to Credit Suisse, have pursued the same common objective as the discussions prior to that date, it would be wrong to consider that Credit Suisse could have been aware or could have reasonably foreseen the continuation of the ‘same’ infringement by the other banks concerned. That would also be the case if it were to be considered that those later discussions were part of the same overall plan and had the same common objective as the earlier discussions. According to Credit Suisse, the grounds put forward by the Commission in the contested decision would certainly make it possible to establish that its trader was aware that the other traders involved, whose personal proximity was well known, had continued their discussions. By contrast, those grounds do not in any way demonstrate that Credit Suisse could have been aware of the anticompetitive nature or purpose of the bilateral discussions between the other traders of the banks concerned. As a preliminary point, it should be recalled that Credit Suisse’s arguments alleging the absence of an overall plan pursuing a single anticompetitive objective were rejected in paragraphs 322 to 362 above. Credit Suisse’s argument cannot therefore succeed in so far as it is based on the absence of an overall plan, in particular after February 2013. As regards the rest of Credit Suisse’s arguments, first, it should be noted that that bank does not dispute that it participated directly in many of the discussions examined by the Commission in the contested decision and which took place in the persistent multilateral chat rooms between June 2010 and February 2013. Secondly, it is clear from paragraph 110 above that the criticisms which Credit Suisse considers to have made of the 25 discussions mentioned in the table annexed to the reply are inadmissible as regards the second part of the first plea in law alleging infringement of Article 101 TFEU by the Commission when it concluded that the ‘price discovery’ constituted a ‘restriction by object’. Moreover, Credit Suisse’s arguments aimed at demonstrating that the discussions of 28 September 2010, 8 February 2012 and 10 January 2013 were not anticompetitive were rejected in paragraphs 269 to 295 above. Thirdly, it is established that Credit Suisse participated directly in or, at the very least, was aware of the existence, prior to February 2013, of anticompetitive bilateral exchanges. The Commission referred to anticompetitive bilateral discussions in non-persistent chat rooms between the Credit Suisse trader and a Deutsche Bank trader on 19 and 20 May 2011 (recitals 294 and 297) and on 17 December 2012 (recital 444), the content of which is not disputed by Credit Suisse. Fourthly, Credit Suisse does not dispute that it was aware that the Deutsche Bank traders had ceased using persistent multilateral chat rooms as the result of their employer’s prohibition, in February 2013, on the use of such chat rooms, and not of any intention on the part of the Deutsche Bank traders to cease participating in the infringement (see in particular recitals 94, 766, 767, 822, 823 and 847). Fifthly, Credit Suisse was directly involved in anticompetitive bilateral exchanges after February 2013 with the other banks concerned. First of all, the Credit Suisse trader took part in anticompetitive bilateral discussions with the Deutsche Bank trader and then the BofA trader on 13 March, 27 August, 25 September and 10 December 2013, as well as on 3 and 5 February, 28 March and 22 July 2014. Next, the Credit Suisse trader participated in anticompetitive bilateral discussions with the Crédit agricole trader on 19 March, 24 May and 9 and 14 August 2013, as well as on 4, 7 and 14 March, 1 May and 4 August 2014. As set out in paragraphs 296 to 308 above, the Credit Suisse trader also engaged in anticompetitive exchanges with the Crédit agricole trader on 12 and 24 March 2015. Until August 2013, the Credit Suisse and Crédit agricole traders continued to use, for their discussions, the persistent multilateral chat room that had been used prior to February 2013 by all participants. Finally, an examination of the bilateral discussions that took place after February 2013 between the Credit Suisse trader and the traders of the other banks concerned shows that that bank participated in the various types of conduct that make up the single and continuous infringement. Sixthly, as Credit Suisse confirmed at the hearing, Credit Suisse does not dispute that its trader knew and was aware of the existence of parallel discussions between the Deutsche Bank and Crédit agricole traders. By contrast, Credit Suisse disputes that its trader was aware of the anticompetitive nature of the discussions between the other two traders. In that regard, as regards the discussions of 25 July and 22 October 2013 and of 9 January 2014, referred to in recitals 808 and 823 of the contested decision, it is true that, as Credit Suisse claims, those discussions alone do not demonstrate that it was aware, through its trader, of the anticompetitive nature of the bilateral exchanges between the Crédit agricole trader and the Deutsche Bank, and then BofA, trader. However, those discussions support the assessment in recital 808 of the contested decision that, in the context of the prohibition on the Deutsche Bank traders to use persistent multilateral chat rooms, the traders continued to keep each other informed by means of a network of bilateral chat rooms and continued to disclose, in a bilateral manner, information obtained from the third trader concerning certain aspects which were usually covered in multilateral chat rooms. During the bilateral discussion of 25 July 2013, the Credit Suisse trader informed the Deutsche Bank trader that the Crédit agricole trader ‘got lifted this am too’. Thus, the Credit Suisse trader revealed, in the context of his bilateral discussion with the Deutsche Bank trader, information that he had obtained in a separate bilateral discussion with the Crédit agricole trader about the latter’s trading activity. That revelation took place in the course of a discussion during which the Deutsche Bank trader indicated: ‘i sold some kfw 09/15 aswell’ then ‘thats me lifting 14.5 and 15 this morning’ and ‘still looking for those if you see any’. For his part, the Credit Suisse trader explained to the Deutsche Bank trader: ‘i shorted those kfw 15s this am to my syndicate actually. probably a bit stupid for a short date, but should be ok’ and then revealed the transaction carried out by the Crédit agricole trader. The spontaneous, natural and frank nature of that conversation shows that it was normal for both the Credit Suisse trader and the Deutsche Bank trader to exchange precise and recent information not only concerning themselves, but also concerning the Crédit agricole trader. The conduct of the Credit Suisse trader during the discussion of 25 July 2013 attests to the climate of mutual trust between the traders of Deutsche Bank, Crédit agricole and Credit Suisse. That conduct therefore constitutes a factor which supports the assessment set out in recital 808 of the contested decision concerning the existence of a network of bilateral discussions in which the three traders discussed their trading activities and therefore supports the fact that the Credit Suisse trader could reasonably have envisaged that the other two traders exchanged sensitive information, in particular concerning their respective trading activities. Furthermore, the fact that the Deutsche Bank trader received identical information from the Crédit agricole trader himself during a bilateral discussion afterwards on the same day, namely 25 July 2013, does not call that assessment into question and corroborates the finding that the traders of the banks concerned were discussing the same obligations and the same subjects in parallel bilateral discussions. As for the discussion of 9 January 2014, during which the Credit Suisse trader informed the Crédit agricole trader that he had just been chatting to the Deutsche Bank trader (‘[was] just chatting to [Deutsche Bank’s trader]’), it does not indeed refer to bilateral discussions relating to trading activities, nor a fortiori to discussions relating to commercially sensitive information or, more generally, to the anticompetitive nature of such discussions. However, it does confirm that the Credit Suisse trader had no hesitation in disclosing to one of the traders the existence of the contacts he had with the third trader. As regards the bilateral discussion between the Crédit agricole trader and the Deutsche Bank trader on 22 October 2013, referred to in paragraph 466 above, it is true that there is no evidence that the Crédit agricole trader approached the Credit Suisse trader following that discussion. However, following his discussion with the Deutsche Bank trader, the Crédit agricole trader wrote: ‘ok will ask him’. The discussion of 22 October 2013 therefore shows that the disclosure of information relating to a third trader, in this case the Credit Suisse trader, was as natural between those two other traders as it was between the Credit Suisse trader and the Deutsche Bank trader, as noted in paragraph 499 above. That discussion therefore supports the assessment in recital 808 of the contested decision concerning the existence of a network of bilateral discussions. Seventhly, it should be noted that, on 11 October 2013, Credit Suisse had an anticompetitive discussion with the Deutsche Bank trader during which the Crédit agricole trader was ‘present’ (see paragraphs 462 and 463 above). That discussion, which Credit Suisse described as ‘multilateral’ and during which the other two traders were present, is further evidence that the Credit Suisse trader could conceive of the existence of anticompetitive bilateral discussions between the other two traders. Thus, because of its participation in the persistent multilateral chat rooms before 25 February 2013, Credit Suisse knew that, before that date, the traders of Deutsche Bank and Crédit agricole were consulting each other. Furthermore, given that, before and after that date, the Credit Suisse trader took part in anticompetitive bilateral discussions with each of those two traders and that the Commission had evidence to show the existence, after 25 February 2013, of a network of bilateral discussions in which the three traders continued to discuss their trading activities, it must be considered that the Credit Suisse trader could, at the very least, reasonably foresee that the other two traders were conducting bilateral discussions which were anticompetitive in nature. In that regard, neither the fact that Credit Suisse did not know in detail the discussions in which its trader had not participated, nor the fact that it was unaware of the existence of some of those discussions, is such as to invalidate the Commission’s finding that it participated in the single infringement and may be held liable for it as a whole (see, by analogy, judgment of 14 December 2006, Raiffeisen Zentralbank Österreich and Others v Commission , T‑259/02 to T‑264/02 and T‑271/02, EU:T:2006:396, paragraph 193). The finding that Credit Suisse could, at the very least, reasonably foresee all the other banks’ conduct is also corroborated by other evidence. First of all, the Commission found that the Credit Suisse trader was part of the restricted circle of traders particularly involved in the infringement at issue. Those traders were associated professionally and personally, including after February 2013, and had been targeted as a group (see recitals 425, 807, 809 and 810 of the contested decision). In that respect, a bilateral discussion between the Crédit agricole trader and the Credit Suisse trader on 8 March 2013, the existence and content of which are not disputed by Credit Suisse, shows that those traders and the Deutsche Bank trader continued to maintain regular contact and organised meetings of a social nature. Similarly, a discussion on 23 May 2013 between the Credit Suisse trader and the Deutsche Bank trader shows that traders of the banks concerned in the infringement were perceived as a ‘gang’ by other traders (see paragraph 478 above). Finally, a discussion between the Credit Suisse trader and the Deutsche Bank trader on 11 October 2013 (recital 530), in which the Crédit agricole trader was present, establishes that the three traders had planned to meet during business trips to New York the following week (see paragraph 393 above). In the context of an analysis of all the evidence, that circumstance is relevant to establishing whether Credit Suisse knew or could reasonably have foreseen the anticompetitive content of the bilateral discussions that took place after February 2013 and until the date on which its participation in the infringement at issue ended, namely 24 March 2015. In the light of all the foregoing, Credit Suisse is not entitled to argue that ‘the evidence gathered by the Commission is insufficient to infer that any of the traders [had] sufficient knowledge of the content of the communications in which he did not participate [and,] in the alternative, that it cannot be held responsible for the bilateral discussions between the other two traders after February 2013’. The second part of Credit Suisse’s second plea in law is therefore rejected. (e) The duration of Crédit agricole’s and Credit Suisse’s participation in the single and continuous infringement at issue In the second part of its second plea in law, Crédit agricole challenges the continuing nature of its participation in the single infringement at issue (see paragraphs 428 to 432 above). By the fourth part of its second plea in law, Credit Suisse alleges errors relating to its participation in the infringement beyond August 2014. (1) The second part of Crédit agricole’s second plea in law, in so far as it seeks to demonstrate that its participation in the infringement found was not continuous In recital 790 of the contested decision, the Commission considered that the banks concerned should be held liable for the entire single and continuous infringement for their respective periods of involvement in it. In recital 842 of that decision, the Commission found that Crédit agricole had participated in the infringement from 10 January 2013 to 24 March 2015, that is to say, a period of 804 days or 2.20 years. Crédit agricole claims that the Commission relies solely on 31 discussions involving it to establish a continuous infringement over a period of 26 months, namely between January 2013 and March 2015, on the understanding that two thirds of those discussions took place in the first 10 months of that period, namely between January and October 2013. Those circumstances call into question the findings that, first, the discussions between the Crédit agricole trader and the traders of the other banks concerned had been ‘constant’ and, secondly, that that bank had participated in a continuous infringement, at least as regards the period after October 2013. More specifically, in the first place, Crédit agricole stresses that a single discussion supports the allegation that it participated in category 1 conduct. According to Crédit agricole, that single discussion is not sufficient to establish coordination in relation to the prices offered to specific counterparties over a period of more than two years. Similarly, with regard to category 2 conduct, the Commission would rely solely on four discussions to demonstrate the existence of an infringement lasting more than two years. In the second place, Crédit agricole’s alleged participation varies significantly over time, even if all 31 discussions identified by the Commission were taken into account. Given the use of chat rooms on the Bloomberg platform and therefore the availability of evidence, that variation would necessarily represent gaps without infringing conduct, rather than periods for which the Commission would have no evidence. Similarly, first, the Commission sought to reverse the burden of proof, relying on the fact that Crédit agricole had not publicly distanced itself from the conduct at issue. Secondly, the Commission wrongly claimed that Crédit agricole was aware of the conduct of the other banks concerned, which that bank disputes by another plea in law. In that regard, it is clear from Crédit agricole’s arguments that the main factor it puts forward to support its allegation is the length of the gaps between the discussions in which its trader took part, in particular if some of those discussions were to be disregarded. In the first place, it should first of all be noted that the Commission did not err as regards the existence of an overall plan pursuing a single anticompetitive objective. Moreover, the identification of five categories of conduct in the contested decision is not such as to call into question the existence of that overall plan (see paragraph 364 above). Accordingly, Crédit agricole cannot be upheld when, for the purposes of finding an alleged interruption in its participation in the infringement at issue, it distinguishes between the categories of conduct identified by the Commission in the contested decision and chooses to disregard certain discussions. In other words, the continuing nature of Crédit agricole’s participation in the single infringement at issue cannot be assessed by examining, category by category, the conduct found against it by the Commission. That continuous nature must be assessed in the light of all the conduct at issue in which that bank’s trader participated, of which he was aware or which he could reasonably have foreseen, whatever category that conduct falls into. Thus, assuming that they are established, which is not the case, the absence of Crédit agricole’s ongoing participation in conduct falling within categories 1 or 2 alone, or its lack of knowledge of the conduct of other participants falling within those categories, are not in themselves likely to call into question the Commission’s findings as to its ongoing participation in the infringement. Next, Crédit agricole is not entitled to rely on gaps on the ground that its third plea in law should be upheld. The third plea in law relied on by Crédit agricole was upheld only in relation to the discussion of 10 January 2013, that is to say, chronologically, the first discussion claimed against it by the Commission. Finally, in so far as Crédit agricole argues that, in order to calculate the gaps between two manifestations of the infringement at issue, discussions falling within some of the five categories of conduct should be disregarded on the grounds that those categories do not constitute a ‘restriction by object’, or because certain discussions constituted ‘legitimate’ activities, it is clear from paragraphs 165 to 268 above that, without prejudice to the question of whether they constitute restrictions by object, all of the conduct classified into five categories by the Commission in which Crédit agricole participated, including the exchanges of information, were anticompetitive in nature. In particular, Crédit agricole’s valid criticisms of certain discussions in which it participated and which fell into the five categories identified by the Commission were rejected in paragraphs 188 to 268 above. It follows that, in order to assess whether the gaps between two manifestations of Crédit agricole’s infringing conduct are likely to demonstrate an interruption in that bank’s participation in the infringement, none of the discussions held against it and in which it directly participated can be excluded, with the exception of the discussion of 10 January 2013. In the second place, in so far as Crédit agricole relies on a five-month gap in the evidence between September 2013 and March 2014, it is important to emphasise that that gap does not take into account the presence of its trader at a discussion on 11 October 2013 and therefore his participation in it (see paragraph 134 above). Thus, as the Commission acknowledges, the maximum gap between two direct participations by the Crédit agricole trader in the infringement at issue is approximately four and a half months and occurred between 11 October 2013 and 4 March 2014. In that regard, first, it should be borne in mind that the question whether or not the duration of those gaps is long enough to constitute an interruption must be assessed in the context of the operation of the infringement at issue (see paragraph 381 above). In the present case, very frequent participation in collusive contacts was not necessary for the operation of the infringement at issue, given that the SSA bond negotiations were independent. The potential success of a given collusive contact did not depend on participation in a series of contacts relating to other negotiations (see paragraphs 384 to 386 above). Secondly, the years 2014 and 2015 were characterised by a reduction in activity on the SSA bond market, beyond the control of the participants in the single infringement at issue, which contributed to a reduction in opportunities to coordinate or exchange commercially sensitive information (see paragraph 390 above). Thirdly, an examination of the discussion of 11 October 2013 and of the discussion of 4 March 2014 shows that they do not contain any element capable of demonstrating that the Crédit agricole trader showed, first, a willingness to put an end to his participation in the single and continuous infringement or, then, a willingness to participate again in that infringement. In particular, the tone and spontaneity of the discussion of 4 March 2014 show a persistent willingness on the part of the Crédit agricole trader to continue to engage in conduct that contributes to the overall plan pursuing a single anticompetitive objective as defined by the Commission in the contested decision. A similar assessment can be made of the discussions that precede each gap in the participation of the Crédit agricole trader and of the discussions that follow those gaps. Fourthly, the change of employer of one of the three main traders involved in the infringement may also have contributed to certain gaps between two manifestations in the participation of the Crédit agricole trader in the infringement at issue (see paragraph 396 above). Fifthly, as is clear from paragraphs 451 to 481 above, the Crédit agricole trader could, at the very least, reasonably have foreseen the existence of anticompetitive discussions between the other two traders even in the absence of any direct participation on his part in those discussions. Thus, first of all, the gaps between two manifestations of Crédit agricole’s participation in the infringement at issue do not lead to the conclusion that that participation was interrupted. Next, Crédit agricole does not dispute that it did not publicly distance itself from the cartel. Finally, Crédit agricole does not invoke any other circumstances which might be capable of demonstrating an interruption in its participation in the infringement at issue, such as, for example, the termination of the duties of the trader who participated in the anticompetitive conduct preceding a gap between two manifestations of that bank’s participation in that infringement and the taking up of duties by a new trader who again contributed to the infringement during a period following that gap. Consequently, having regard, first, to the total duration of Crédit agricole’s participation, which exceeded two years, secondly, to the evidence available to that bank concerning the context in which the cartel operated and the gaps between two manifestations of the Crédit agricole trader’s participation in the infringement found, thirdly, to the absence of any public distancing on the part of Crédit agricole and, fourthly, to the absence of any other circumstances relied on by that bank to demonstrate an interruption in its participation in the infringement at issue, it was without reversing the burden of proof that the Commission found, correctly, that Crédit agricole’s participation in the infringement at issue had not been interrupted between 11 January 2013 - a day whose discussion is not regularly contested by Crédit agricole - and 24 March 2015 (paragraphs 144 and 447 above). The second part of Crédit agricole’s second plea in law is therefore rejected in so far as it seeks to demonstrate that its participation in the single infringement found was not continuous. (2) The fourth part of Credit Suisse’s second plea in law, alleging that the Commission failed to prove that it participated in any unlawful conduct after August 2014 In recital 842 of the contested decision, the Commission found that Credit Suisse had participated in the single and continuous infringement at issue during the period from 21 June 2010 to 24 March 2015, that is to say, a period of 1 738 days or 4.75 years. Furthermore, in recitals 846 to 849 of the contested decision, the Commission rejected Credit Suisse’s argument that it had not proved that that bank had participated in the infringement between August 2014 and March 2015. By the fourth part of its second plea in law, Credit Suisse claims that the Commission has not demonstrated its participation in the infringement at issue beyond August 2014. In that regard, the contested decision refers to only six discussions after that date and uses them to extend that participation until 24 March 2015. Credit Suisse did not take part in the first four discussions and the bank disputes the anticompetitive nature of the last two, namely the discussion of 12 March 2015 between its trader and the Crédit agricole trader and the email sent by its trader to the Crédit agricole trader on 24 March 2015. In that regard, first, Credit Suisse’s argument by which that bank disputed its knowledge or the fact that it could reasonably have foreseen the anticompetitive bilateral discussions involving other banks after February 2013 and until the end of the period of its participation in the infringement retained in the contested decision was rejected in paragraphs 482 to 513 above. Secondly, and more importantly, it is clear from paragraphs 296 to 308 above that Credit Suisse has not demonstrated the existence of an error committed by the Commission when it considered that the discussion of 12 March 2015 and the email of 24 March 2015 were anticompetitive in nature. Credit Suisse is therefore not entitled to argue that the Commission has not demonstrated its involvement in the infringement at issue beyond August 2014. The fourth part of the second plea in law put forward by Credit Suisse is therefore rejected. (f) Conclusion on Crédit agricole’s second plea in law and Credit Suisse’s second plea in law In the light of the foregoing, the second plea in law relied on by Credit Suisse is rejected. As regards Crédit agricole, it is clear from an examination of the second plea in law put forward by that bank that it has not shown that the Commission committed an error of assessment when it considered that it had participated in a single and continuous infringement until 24 March 2015. By contrast, it is important to remember that the Commission could not validly hold against Crédit agricole the discussion of 10 January 2013, which took place before its trader’s first connection to the chat rooms concerned. It follows that the Commission erred when it considered, in Article 1(c) of the contested decision, that Crédit agricole had participated in a single and continuous infringement from 10 January 2013 and not from 11 January 2013. That bank’s second plea in law must therefore be upheld to that extent alone. 3. Crédit agricole’s first plea in law and Credit Suisse’s first plea in law, alleging errors in the classification of the conduct at issue as a restricti on by object In recitals 739 to 749 of the contested decision, the Commission concluded that the conduct at issue had an anticompetitive object, stressing in particular the following: ‘(747) By means of all of the above practices, the participating traders knowingly substituted collusion for normal competition, to the extent that at times they acted as if they were trading the same positions or “book”. Regardless of the form the collusion might take in a specific instance, however, the overriding objective was to benefit their own revenues. The traders sought to achieve this by coordinating on prices, exchanging sensitive information on trading activities, aligning trading and pricing strategies and coordinating trading activity. In doing so they reduced uncertainty and risk, made trades on advantageous terms to themselves, shared markets and clients and maintained a competitive advantage vis-a-vis clients and other market participants. (748) The participating traders did so in a market (the market for … SSA bonds) that should, as other sophisticated debt markets, normally be characterised by competitive pricing and other trading conditions. The participating banks were established market players in an industry designed to facilitate the efficient and competitive issuance and trading of debt capital. (749) Accordingly, it is concluded that the agreements and/or concerted practices set out in Section 4.2 have as their object the restriction and/or distortion of competition within the meaning of Article 101(1) [TFEU] and Article 53(1) of the EEA Agreement.’ In the context of their respective first pleas in law, Crédit agricole and Credit Suisse criticise the Commission for wrongly classifying the conduct at issue as a ‘restriction by object’. (a) Preliminary observations on the subject and merits of the applicants’ criticisms (1) Crédit agricole’s criticisms In the context of its first plea in law, which is divided into two parts, Crédit agricole claims that the Commission did not provide sufficient evidence and reasons for its participation in a restriction by object. In its first part, Crédit agricole relies on ‘errors of law and fact in finding that the categories relating to exchanges of information constitute restrictions by object liable to form part of the single and continuous infringement [found]’. To that end, it argues, first, that the exchanges of information covered by category 3 (‘Exchange of current, or forward-looking commercially sensitive information on their trading activities and trade flows in the secondary market’) did not present a sufficient degree of harmfulness to competition to qualify as a ‘restriction by object’ and, secondly, that the conduct covered by category 4 (‘Exchange, confirmation and alignment of trading and pricing strategies’) did not constitute exchange, confirmation and alignment on trading and pricing strategies but, in fact, exchanges of information which did not present such a sufficient degree of harmfulness either. In the second part of its first plea in law, Crédit agricole complains that the Commission ‘failed to carry out the analysis required to support a finding of a restriction by object in relation to any of the categories of coordinated conduct [found]’, namely categories 1 (‘Coordination on prices quoted to specific counterparties’), 2 (‘Coordination on prices to show to the market generally’) and 5 (‘Coordination of trading activity’). At the outset, it should be noted that Crédit agricole’s arguments relating to the absence of a single and continuous infringement and to its lack of participation in that infringement were rejected in paragraph 550 above, subject to the date on which the participation began, and therefore do not need to be examined in the context of the present plea in law. It should also be noted that, in the arguments summarised in paragraphs 555 and 557 above, Crédit agricole criticises the Commission for finding that each of the five categories referred to in recital 613 of the contested decision had to be classified as a ‘restriction by object’, even though the conduct in categories 3 and 4 was not sufficiently harmful to competition and that, as regards the conduct in categories 1, 2 and 5, it had not carried out the requisite analysis. It has already been noted in paragraph 87 above that that argument is based on an erroneous reading of the contested decision and must therefore be rejected as unfounded. In addition, it was noted in paragraphs 106 and 107 above that, in its first plea in law, Crédit agricole challenged in a specific, precise and substantiated manner only the discussions of 18 and 31 January, 15 February, 19 and 21 March, 24 May, 3 June and 10 and 25 July 2013, as well as 6 August 2014 and 12 March 2015, and that, as a result, the assessments made by the Commission concerning the discussions other than those ones are now definitively established. However, given that the Court cannot confine itself to examining the arguments raised by the parties solely in the context of the plea, branch or complaint in support of which they were formally raised, but must also assess them in support of another plea, branch or complaint where those arguments are also relevant in support thereof (see, to that effect, judgment of 21 September 2023, Fachverband Spielhallen and LM v Commission , C‑831/21 P, EU:C:2023:686 , paragraphs 46 to 49), Crédit agricole’s first plea in law must be interpreted as meaning that, first, the Commission failed to assess the economic context of the infringement at issue as a whole or assessed it incorrectly and, secondly, the Commission wrongly concluded that the 11 discussions referred to in paragraph 561 above were capable of supporting the conclusion that the single and continuous infringement found by that institution had an anticompetitive object. (2) Credit Suisse’s criticisms In its first plea in law, which is divided into three parts, Credit Suisse claims that ‘the Commission infringed Article 101 TFEU and gave insufficient reasons for its finding that the bank had engaged in conduct which had the object of restricting or distorting competition’. In the first part of its first plea in law, which is submitted as its main plea and consists of four points, Credit Suisse criticises the Commission on the ground that, for the purposes of assessing the discussions as a whole, it ‘did not take sufficient account of the factual and legal context of the SSA bond market and did not discharge the burden of proving that the conduct at issue restricted competition by object’. In the second part of its first plea in law, submitted in the alternative and consisting of 24 points, Credit Suisse criticises the Commission for having ‘concluded … that price discovery communications restricted competition by object’, essentially on the ground that such communications ‘could be pro-competitive’. In the third part of its first plea in law, Credit Suisse claims that, when examining the pro-competitive effects of that price discovery, the Commission infringed the rules of evidence and its rights of defence by applying to the conduct at issue not the rules applicable to proof of the anticompetitive object of conduct, but those applicable to proof of the ancillary nature of a restriction, within the meaning of paragraph 89 of the judgment of 11 September 2014, MasterCard and Others v Commission (C‑382/12 P, EU:C:2014:2201 ). With regard to that argument, it should be noted, first, that, in recital 749 of the contested decision, the Commission found that the conduct at issue, described in Section 4.2 of that decision, was to be classified as a ‘restriction by object’, after finding, in recital 104, that the various categories of conduct which it selected were described only for analytical purposes and that they were linked and often overlapped. Accordingly, that classification was not applied either to each of the categories referred to or to each of the forms of conduct at issue. There is therefore no reason to qualify the 25 discussions in the category identified by Credit Suisse under the heading ‘price discovery’, the dates of which are set out in Annex C.1 to Credit Suisse’s reply. Secondly, as is apparent from paragraph 110 above, unless they are formulated even summarily in the application or reply, the criticisms directed specifically at the 25 discussions grouped together by Credit Suisse in the category entitled ‘price discovery’ are inadmissible. It follows that the second part of Credit Suisse’s first plea in law, alleging the erroneous classification of the price discovery as a ‘restriction by object’, must be rejected. That being so, the arguments raised by Credit Suisse in support of the second part of its first plea in law and relating to the alleged pro-competitive effect of the discussions on price discovery in the context of the first part of the first plea in law must be examined in so far as they support its criticisms that the Commission made errors in classifying the conduct at issue, taken as a whole, as a ‘restriction by object’, in particular because insufficient account was taken of the legal and economic context of the infringement at issue. In so doing, account must be taken of arguments capable of calling into question the classification of the conduct at issue as a ‘restriction by object’, even though they were formally put forward in support of the second part of the first plea in law, which was rejected, but are equally relevant to the assessment of the first part of that plea in law (see, by analogy, judgment of 21 September 2023, Fachverband Spielhallen and LM v Commission , C‑831/21 P, EU:C:2023:686 , paragraph 48). (b) The classification of the conduct at issue constitutes a ‘restriction by object’ (1) Preliminary observations It should be borne in mind that, in order to fall within the prohibition in principle laid down in Article 101(1) TFEU, a concerted practice must have as its ‘object or effect’ the prevention, restriction or distortion to an appreciable extent of competition within the internal market (judgment of 30 January 2020, Generics (UK) and Others , C‑307/18, EU:C:2020:52 , paragraph 62). It follows that that provision, as interpreted by the Court of Justice, draws a clear distinction between the concept of ‘restriction by object’ and that of ‘restriction by effect’, each being subject to different rules of evidence (judgment of 30 January 2020, Generics (UK) and Others , C‑307/18, EU:C:2020:52 , paragraph 63). Thus, as regards practices classified as a ‘restriction by object’, there is no need to investigate or, a fortiori, to demonstrate their effects on competition in order to classify them as a ‘restriction of competition’ within the meaning of Article 101(1) TFEU, in so far as experience shows that such behaviour leads to falls in production and price increases, resulting in poor allocation of resources to the detriment, in particular, of consumers. Concerning such practices, all that is required is the demonstration that they can in fact be classified as ‘restrictions by object’, though mere unsubstantiated allegations are not however sufficient (judgment of 30 January 2020, Generics (UK) and Others , C‑307/18, EU:C:2020:52 , paragraphs 64 and 65). By contrast, where the anticompetitive object of an agreement, a decision by an association of undertakings or a concerted practice is not established, it is necessary to examine its effects in order to prove that competition has in fact been prevented or restricted or distorted to an appreciable extent (see judgment of 30 January 2020, Generics (UK) and Others , C‑307/18, EU:C:2020:52 , paragraph 66 and the case-law cited). It is also clear from the Court of Justice’s case-law, noted, in essence, in recitals 623 to 626 of the contested decision, that the concept of ‘restriction by object’ must be interpreted strictly and may be applied only to some concerted practices between undertakings which reveal, in themselves and having regard to the content of their provisions, their objectives, and the economic and legal context of which they form part, a sufficient degree of harm to competition for it to be found that there is no need to examine their effects, since some forms of coordination between undertakings can be regarded, by their very nature, as being harmful to the proper functioning of normal competition. When determining that context, it is necessary to take into consideration the nature of the goods or services affected, as well as the real conditions of the functioning and structure of the market or markets in question (see, to that effect, judgment of 30 January 2020, Generics (UK) and Others , C‑307/18, EU:C:2020:52 , paragraphs 67 and 68 and the case-law cited). (2) The degree of harmfulness to competition required to classify conduct as a ‘restriction by object’ Crédit agricole considers that the case-law sets an ‘extremely high bar’ for concluding that an exchange of information can be classified as a ‘restriction by object’, which would be confirmed by paragraphs 72 to 74 of the Guidelines on the applicability of Article 101 [TFEU] to horizontal co-operation agreements (OJ 2011 C 11, p. 1) and by the judgment of 24 September 2019, HSBC Holdings and Others v Commission (T‑105/17, EU:T:2019:675 , paragraph 184). Credit Suisse also suggests that conduct can only be classified as a ‘restriction by object’ if experience and economic theory ‘clearly’ show that it is harmful to competition. In that regard, it follows from paragraphs 573 to 577 above 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 - whether in the form of exchanges of information or other types of anticompetitive conduct - did not present an ‘extremely high bar’ for harmfulness to competition, as Crédit agricole argues, but only a sufficient degree of harmfulness to competition, as the Court pointed out in its judgment of 12 January 2023, HSBC Holdings and Others v Commission (C‑883/19 P, EU:C:2023:11 , paragraph 106). That conclusion that the Commission does not have to establish an ‘extremely high bar’ for harmfulness to competition in order to classify conduct as a ‘restriction by object’ cannot be overturned by paragraphs 72 to 74 of the Guidelines on the applicability of Article 101 [TFEU] to horizontal co-operation agreements. The fact that those guidelines state that ‘it is less likely that information exchanges concerning future intentions are made for pro-competitive reasons than exchanges of actual data’ could certainly lead to the conclusion that exchanges of information relating to future conduct are a priori more problematic than exchanges of information relating to current data. However, under no circumstances can those guidelines lead to the conclusion that exchanges of current information cannot be qualified as a ‘restriction by object’ or that such a qualification requires a higher level of proof of their harmfulness to competition than that retained by the case-law referred to in paragraph 580 above. Similarly, Credit Suisse is wrong to argue that experience and economic theory should ‘clearly’ establish the harmfulness to competition and, therefore, the inherently anticompetitive nature of the conduct which the Commission describes as a ‘restriction by object’. First, notwithstanding paragraph 76 of the judgment of 2 April 2020, Budapest Bank and Others (C‑228/18, EU:C:2020:265 ), referred to by that bank, the Court of Justice expressly held, subsequent to that judgment, that it was by no means necessary for the same type of conduct as that which was to be classified by the Commission in a given case to have already been condemned by it in order for it to be regarded as restrictive of competition by object (see, to that effect, judgment of 25 March 2021, Xellia Pharmaceuticals and Alpharma v Commission , C‑611/16 P, EU:C:2021:245 , paragraph 119). Secondly, as regards the argument based on economic theory, Credit Suisse does not provide any evidence in support thereof, nor does it define its precise scope. Moreover, in its reply, it stated that it did not claim that the assertions in the economic literature were decisive in establishing whether conduct had an anticompetitive object, nor did it seek to impose on the Commission the burden of obtaining the economic literature that would lead to the conclusion that conduct restricted competition by object. Furthermore, there is nothing in the case-law to suggest that, as Crédit agricole argues, the sufficient degree of harmfulness of conduct with regard to competition to enable it to be classified as a ‘restriction by object’ should be assessed individually and separately for each undertaking taking part in that conduct. On the contrary, that assessment, which determines the evidential regime applicable to the conduct at issue as a whole (see, to that effect, judgment of 30 January 2020, Generics (UK) and Others , C‑307/18, EU:C:2020:52 , paragraph 63), must be made in the light of the objective characteristics of that conduct and without regard to the particular situation of each undertaking which participated in it. Thus, as Crédit agricole admitted at the hearing, the minor role of 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 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, such as those of 18 or 31 January 2013, or on the fact that the Commission cited only one example of category 1 discussions, in order to challenge the classification of the conduct at issue as a ‘restriction by object’. (3) The Commission’s assessment of the conduct at issue As is apparent from paragraphs 554 to 557 and 563 to 566 above, Crédit agricole and Credit Suisse, in essence, criticise the Commission for errors in assessing the economic context of the conduct at issue, in assessing whether it was harmful to competition and in assessing whether it was, in essence, ‘justified’. (i) Criticism of the Commission’s assessment of the economic context of the conduct at issue As part of their criticisms alleging the erroneous classification of the conduct at issue as a ‘restriction by object’, Crédit agricole and Credit Suisse criticise the Commission, together or separately, both for failing to state reasons and for erroneously stating reasons for the contested decision as regards consideration of the economic context of the infringement at issue. – Crédit agricole’s criticism, alleging that it failed to assess the economic context of the infringement at issue By the second part of its first plea in law, Crédit agricole criticises the Commission, inter alia, for a ‘failure to conduct the necessary analysis to support a finding of an infringement by object’. That argument, which must be understood as criticising the Commission for breach of its obligation to state reasons as regards the economic context of the infringement at issue, is unfounded. It is clear from recitals 3 to 65 of the contested decision that the Commission took care to present in detail the products in question, the participants and the operation of the secondary market in SSA bonds and the trading practices on that market. In particular, it set out the specific role of market makers, whose functions it noted in recital 19 of that decision. Furthermore, in response to the arguments put forward by the banks concerned in the course of the administrative procedure, the Commission provided, in recitals 690 to 734 of the contested decision, the reasons why neither the market-making function performed on a voluntary basis by those banks nor the usefulness of price discovery called into question its assessment of the anticompetitive object of the conduct at issue. In that context, the Commission noted, in recitals 690 to 703 of the contested decision, that the banks concerned had an important market-making function, which had justified it not criticising them for discussions exclusively intended for the negotiation and execution of bilateral transactions and the offsetting of risks on the inter-traders market, as it had indicated in recital 662 of that decision (see, by analogy, judgment of 24 September 2019, HSBC Holdings and Others v Commission , T‑105/17, EU:T:2019:675 , paragraph 152). However, the Commission rejected, with reasons, the arguments based on the exercise of that market-making function, noting, in recital 695 of the contested decision, that it was clear that the role of the banks concerned as traders did not consist simply in ‘promot[ing] liquidity to support the bank’s clients wishing to trade SSA bonds’, as the applicants had argued in the course of the administrative procedure. Equally reasoned, it indicated that the discussions at issue went beyond the legitimate search for liquidity for risk mitigation purposes, in particular in that they led to traders acting in concert without the knowledge of their clients, which reduced the latter’s possibilities of benefiting from prices set autonomously and competitively. Similarly, traders worked together, sharing knowledge of their clients’ approach, in order to carry out transactions with their clients that corresponded to their mutual interests (recital 703 of the contested decision). In recitals 704 to 713 of the contested decision, the Commission also explained why the arguments alleging the need for price discovery, arising from an asymmetry of information in favour of certain investors and to the detriment of the banks concerned, had to be rejected. In recital 709 of the contested decision, it thus noted that SSA bond traders were specialised in a category of securities and could be expected to be well informed about relevant external events. It also noted, in recital 710 of that decision, that the traders concerned did not merely exchange information in order to discourage smart-money clients, but that they exchanged commercially sensitive information on all types of client. The Commission also explained why it had refused to uphold the allegation that the exchanges of information in question had pro-competitive effects (recital 713 of the contested decision). Finally, in recitals 714 to 719 of the contested decision, the Commission rejected Crédit agricole’s argument that the data exchanged were public or rapidly obsolete, noting in particular the usefulness of information relating to recent transactions for assessing market trends and future trading opportunities. In view of the foregoing, Crédit agricole cannot validly maintain that the Commission failed to fulfil its obligation to state reasons as regards the description of the economic context of the infringement at issue. Accordingly, the second part of its first plea in law must be rejected as unfounded. – Crédit agricole’s and Credit Suisse’s criticisms, alleging the erroneous assessment of the economic context of the infringement at issue In the context of the first part of their respective first pleas in law, first, Crédit agricole and Credit Suisse complain that the Commission took insufficient account of the characteristics of the SSA bond market. In particular, the Commission did not examine the oligopolistic or, on the contrary, atomistic nature of that market, even though the concentration of the market concerned is a relevant factor in assessing whether exchanges of information are harmful to competition. Secondly, the Commission wrongly assessed the operation of the secondary market in SSA bonds, essentially in that it did not take due account of the market-making function performed by the banks concerned, which led it wrongly to consider that some of the conduct at issue could be regarded as corroborating the anticompetitive object of the single and continuous infringement at issue. Thus, the Commission did not take proper account of the fact that the provision of immediacy services by market makers implied that they traded with each other and constantly and thoroughly researched prices (‘the market colour’) and sources of liquidity, but also of the fact that there was a degree of cooperation between them and regular flows of commercially sensitive information. In that respect, Crédit agricole argues that ‘trader[s] need to understand how prices have moved in the past in order to know what [they need] to do to win business’. Thirdly, according to Credit Suisse, the Commission failed to fulfil its obligation to demonstrate that particular circumstances were not capable of giving rise to doubt as to the alleged harmfulness of the agreement concerned, as required by point 48 of the Opinion of Advocate General Bobek in Budapest Bank and Others (C‑228/18, EU:C:2019:678 ). As regards, first, the criticism that insufficient account is taken of the characteristics of the SSA bond market, the Commission submits that the Court of Justice has already had occasion to hold that, in the case of agreements which constitute particularly serious infringements of competition, the analysis of the economic and legal context of which the practice forms part may be limited to what is strictly necessary in order to establish the existence of a restriction of competition by object (see judgment of 27 April 2017, FSL and Others v Commission , C‑469/15 P, EU:C:2017:308 , paragraph 107 and the case-law cited). Nevertheless, consideration of all the factors referred to in paragraph 577 above must, in any event, show the precise reasons why the conduct at issue reveals a sufficient degree of harm to competition, such as to justify a finding that it has as its object the prevention, restriction or distortion of competition (see, to that effect, judgments of 11 September 2014, CB v Commission , C‑67/13 P, EU:C:2014:2204 , paragraph 69, and of 21 December 2023, European Superleague Company , C‑333/21, EU:C:2023:1011 , paragraph 168). Moreover, in a context such as that of the SSA bond market, the Commission cannot rely on the case-law referred to in paragraph 606 above. That is in particular because, in a complex market such as the one at issue in the present case, characterised by a constant and substantial flow of information between the various players involved, determining whether the conduct at issue is sufficiently harmful requires a detailed analysis of the discussions between the traders of the banks concerned. In that regard, the Commission itself acknowledged, in recitals 46 and 662 of the contested decision, that certain discussions - which had a social purpose or were intended to explore the possibility of bilateral transactions or to discuss the market colour, which was already known to all - were not covered by that decision. However, it must be noted, as pointed out in paragraphs 594 to 599 above, that the Commission took into account, in particular, the nature of the products in question and the functions carried out by the banks concerned. Therefore, contrary to what Credit Suisse maintains, that institution did not confine itself to assuming, by referring to the exchange of information on other markets or financial markets other than the SSA bond market, that the conduct at issue considerably reduced the normal uncertainties of the SSA bond market to the benefit of the banks concerned and to the detriment of other market operators. While it is true that, in the contested decision, the Commission did not expressly take account of the concentration of the SSA bond market, it cannot validly be criticised for such an omission. Apart from the fact that the applicants do not state that they raised, in the course of the administrative procedure, the need to take account of that factor for the purposes of classifying the conduct at issue as a ‘restriction by object’, the criticisms alleging that failure to analyse the concentration of the SSA bond market are based essentially on general references to previous case-law and thus do not highlight the importance that such an analysis would have had in the present case. Thus, the applicants have not shown to what extent that analysis was relevant to the conduct of which they are accused, namely price coordination, market or client sharing or exchanges of commercially sensitive information. As regards the exchanges of commercially sensitive information, it must also be noted that, although the Court of Justice was initially able to take into consideration the oligopolistic nature of the market on which the exchanges of commercially sensitive information took place (see, to that effect, judgment of 28 May 1998, Deere v Commission , C‑7/95 P, EU:C:1998:256, paragraphs 75 and 88 to 90), it is apparent, in particular, from the judgment of 19 March 2015, Dole Food and Dole Fresh Fruit Europe v Commission (C‑286/13 P, EU:C:2015:184 ), that the EU Courts were able to recognise the anticompetitive object of such exchanges, without taking into account the oligopolistic nature of the product market on which they took place (see, to that effect, judgment of 29 September 2021, Nichicon Corporation v Commission , T‑342/18, EU:T:2021:635 , paragraphs 105 to 112 (not published)), or even in the absence of such market concentration (see, to that effect, judgment of 14 March 2013, Dole Food and Dole Germany v Commission , T‑588/08, EU:T:2013:130, paragraph 334). With regard, secondly, to the criticism alleging the erroneous assessment of the operation of the SSA bond market, it should be noted that the arguments put forward by Crédit agricole and Credit Suisse do not call into question the findings made by the Commission in recitals 19 and 690 to 697 of the contested decision, according to which the banks concerned performed the function of market maker on a voluntary basis, without being bound by the obligations specific to designated market makers, and according to which their role did not consist solely in promoting liquidity to support their clients wishing to trade in SSA bonds. It is clear from the six discussions referred to by the Commission in support of those findings (recitals 693 and 696 of the contested decision) and not contested by the applicants that, while those banks certainly performed a market-making function, their trading activity was not limited to that function alone and was also aimed at maximising their profits. Moreover, despite the fact that the banks concerned had no official market-making function, the Commission nevertheless took care to take account, in the contested decision, of the exercise of that function. The Commission thus distinguished between trading activities between market makers and end-investors and trading activities between market makers, in particular for liquidity-sourcing purposes. In that sense, it stated, in recitals 42 and 43 of the contested decision, that the present case did not concern discussions between the operators concerned in the normal course of their business, relating to matters such as the provision of necessary information and intended to explore the possibilities of transactions between them as potential counterparties or clients, or comments on the market colour, which it confirmed in recitals 661 to 663 and 699 of the contested decision, noting that it did not call into question the information exchanged in the context of contractual negotiations between the traders of the banks concerned. Accordingly, the Commission was right to consider, in particular in recitals 649 and 650 of the contested decision, that Crédit agricole and Credit Suisse could not usefully rely on their role as market makers, since their conduct went beyond what was normally expected of such market makers. As regards, thirdly, the criticism drawn from point 48 of the Opinion of Advocate General Bobek in Budapest Bank and Others (C‑228/18, EU:C:2019:678 ), it must be noted that, as follows in particular from paragraphs 616 to 621 above, the Commission rightly and in any event rejected the applicants’ arguments based on the specific circumstances of the SSA bond market and, in particular, of the market-making functions of the banks concerned. Consequently, Crédit agricole’s and Credit Suisse’s arguments based on an erroneous assessment of the SSA bond market - and in particular of their function as market makers - must be rejected as unfounded. However, notwithstanding the fact that, in the contested decision, the Commission did not find as many infringements as there are categories referred to in recitals 103 and 613 of the contested decision (paragraphs 87, 364 and 445 above), that conclusion does not prejudge the fact that, for each type of conducts at issue, or even for each disputed discussion, the Commission may have made an erroneous assessment of whether they were sufficiently harmful to competition to contribute to the classification of the conduct at issue as a whole as a ‘restriction by object’, which will be examined below. (ii) Crédit agricole’s and Credit Suisse’s criticism, alleging that the conduct at issue is not sufficiently harmful to competition In the context of their criticism alleging that the conduct at issue was wrongly classified as a ‘restriction by object’, Crédit agricole and Credit Suisse criticise the Commission for wrongly considering that it was sufficiently harmful to competition to be classified as a ‘restriction by object’. According to Crédit agricole, the exchanges of information classified in categories 3 and 4 do not relate to future prices or forward-looking information, but to past prices, potential transactions or non-price concerns relating to specific clients. Moreover, they are not part of systematic exchanges. Such exchanges of information - such as those of 19 March, 24 May and 25 July 2013, classified in category 3, and those of 19 March and 3 June 2013 and 6 August 2014, classified in category 4 - could not therefore be taken into account for the purposes of classifying the conduct at issue as a ‘restriction by object’, in so far as ‘there is an alternative interpretation of them, clearly plausible, to that of the Commission’. In addition, the fact that the Commission considered that the exchanges of information in question would generate an asymmetry of information to the benefit of the banks concerned is not relevant, since the SSA bond market, by its very nature, suffers from an asymmetry of information. Finally, the Commission did not examine whether the exchanges of information had reduced or eliminated the degree of uncertainty about the operation of the SSA bond market. Nor could the conduct classified in categories 1, 2 and 5 be regarded as sufficiently harmful to competition to be taken into account for the purposes of classifying the conduct at issue as a ‘restriction by object’, in so far as the discussions of 18 and 31 January, 15 February, 21 March and 10 July 2013 and 12 March 2015 do not reveal any coordination between the banks concerned. According to Credit Suisse, it would not be open to criticism for market makers to ‘kill a price’ posted on a broker’s screen in order to trade directly with each other, as in the discussions of 28 September 2010 and 10 January 2013. As is apparent from paragraph 178 of the judgment of 24 September 2019, HSBC Holdings and Others v Commission (T‑105/17, EU:T:2019:675 ), it would also not be open to criticism for market makers to exchange information on their inventory positions (long or short), in so far as that would make it easier to find opportunities for transactions between traders. As a preliminary point, it should be noted that the General Court has already found, in paragraphs 185 to 310 above, that the Commission was right to find that the conduct was anticompetitive. Consequently, in the context of the examination of the applicants’ criticisms referred to in paragraphs 626 to 628 above, it is only necessary to assess whether that conduct is sufficiently harmful to competition for it, taken as a whole, to be classified as a ‘restriction by object’. To that end, it should be noted, first, that the Commission accused the banks concerned of a single and continuous infringement consisting of more than 200 discussions, referred to in Annex 1 to the Statement of Objections, to which footnote 118 to the contested decision refers, of which more than 120 are analysed in the body of that decision and of which, in their respective first pleas in law, only 11 were properly contested by Crédit agricole (see paragraphs 100 to 107 and 625 and 627 above) and 3 by Credit Suisse (see paragraphs 108 to 113 and 628 above). Secondly, the Commission qualified those discussions as a whole as a ‘restriction by object’, without using that qualification either for each of those discussions or for each of the five categories in which they were classified. In recitals 614 to 616 of the contested decision, the Commission nevertheless considered, for the purposes of illustration, that the discussions classified in categories 1 and 2 revealed ‘coordination on prices’ between the banks concerned, that those classified in categories 3 and 4 revealed ‘disclosure of sensitive information on a trader’s own motion or upon request in order to be used as intelligence by other traders on trading activities and trading and pricing strategies’, and that the discussions classified in category 5 revealed ‘coordination of trading activity’. It follows that it is necessary to determine whether the criticisms of Crédit agricole and Credit Suisse, examined in conjunction with the 14 discussions validly criticised by them, are such as to call into question the Commission’s classification as a ‘restriction by object’, which is already supported by the assessments relating to the more than 120 discussions which were not contested or were not properly contested by those banks and which must therefore be regarded as definitively established, as must the conclusions which the Commission drew from them. At the outset, it should be pointed out, as already noted in paragraph 186 above, that, where the Commission relies on documentary evidence in establishing an infringement of Article 101 TFEU, it is incumbent on the undertakings concerned not merely to present a plausible alternative to the Commission’s view, but to raise the insufficiency of the evidence relied on in the contested decision in order to establish the existence of the infringement. As regards, as a first step, price coordination conduct - whether carried out in relation to specific counterparties or to the market in general - and coordination of trading activity, it is settled case-law that such conduct is sufficiently harmful to competition to be considered, by its very nature, to be detrimental to the proper functioning of normal competition. The former conduct referred to in the preceding paragraph can be analysed as horizontal price-fixing between competitors which may be considered so likely to have negative effects, in particular on the price, quantity or quality of the goods and services, that it may be considered redundant, for the purposes of applying Article 101(1) TFEU, to prove that it has actual effects on the market (see judgment of 2 April 2020, Budapest Bank and Others , C‑228/18, EU:C:2020:265 , paragraph 36 and the case-law cited). As the Commission rightly pointed out in recitals 645, 648 and 653 of the contested decision, the banks concerned thus agreed on the prices which each of them would offer to a client, either by fixing identical prices for that client or by agreeing that one of them would offer a more competitive price than the other. They also jointly discussed and formulated their prices for the bonds they would present to the market through their sales offices or brokers, which took the form of submitting bids or offers to brokers at coordinated prices, or sending axe sheets or comps lists, namely lists produced by a trader in connection with a new bond issue, showing the prices of comparable bonds with similar characteristics (maturity, yield) to those of the new issue (recital 54 of the contested decision), to their respective sales offices with coordinated prices. The second forms of conduct referred to in paragraph 636 above can be analysed as market or client allocation conduct which in themselves have an object restrictive of competition (judgment of 19 December 2013, Siemens and Others v Commission , C‑239/11 P, C‑489/11 P and C‑498/11 P, not published, EU:C:2013:866 , paragraph 218), as the Commission noted, in essence, in recitals 646 and 681 of the contested decision. Such conduct led the banks concerned to refrain from making offers to buy or sell, to withdraw (kill) offers to buy or sell from the market where there was a risk that they would compete or interfere with each other, or to share transactions and merge or reduce their respective positions in order to meet the demand of a specific client (as disclosed between them), which, in particular, limited the bargaining power of those banks’ clients, as the Commission pointed out, in essence, in recital 703 of the contested decision. Furthermore, Crédit agricole and Credit Suisse cannot validly claim that the discussions of 28 September 2010, of 10, 18 and 31 January, 15 February, 21 March and 10 July 2013 and of 12 March 2015 were wrongly taken into account for the purposes of classifying the conduct at issue, as a whole, as a ‘restriction by object’. As regards the discussions of 18 and 31 January 2013, it has already been noted, in paragraphs 188 to 200 above, that the Commission did not commit any error as to the anticompetitive nature of those exchanges. Moreover, the arguments put forward by Crédit agricole do not, in essence, concern the assessment of the particular harmfulness to competition of the conduct at issue. As regards the discussions of 28 September 2010 and 10 January 2013, and as is apparent from paragraphs 269 to 278 and 287 to 295 above, they reveal, at the very least, coordination of the trading activity of Credit Suisse and Deutsche Bank traders - in that one offered to withdraw the price posted on a broker’s screen while the latter was trading with a third party - and without those banks being able to show that the evidence adduced by the Commission as to the anticompetitive nature of those discussions was insufficient. As regards the discussion of 15 February 2013, and as is clear from paragraphs 201 to 208 above, it reveals at the very least coordination on prices, whereby Crédit agricole and Deutsche Bank jointly discussed and formulated their prices for bonds that they would present to a client, and this notwithstanding the fact that a discussion, which Crédit agricole considers to be similar, would not have been reproached to a third-party bank. As regards the discussion of 21 March 2013, and as is clear from paragraphs 217 to 227 above, it certainly cannot be considered to reflect coordination on prices. However, Crédit agricole does not claim or, a fortiori, demonstrate that the exchanges which took place during that discussion were wrongly considered to constitute exchanges of commercially sensitive information which were sufficiently harmful to competition to contribute to the classification of the conduct at issue as a ‘restriction by object’ (see, by analogy, judgment of 12 January 2023, HSBC Holdings and Others v Commission , C‑883/19 P, EU:C:2023:11 , paragraph 151). As regards the discussion of 10 July 2013, and as is clear from paragraphs 236 to 243 above, it reveals at the very least coordination on prices as well as coordination of the trading activity of Crédit agricole and Deutsche Bank, having led the trader of the former bank to indicate to the trader of the latter bank that he was going to make a bid at the same price as the latter had previously communicated to him. Finally, as regards the discussion of 12 March 2015, and as is clear from paragraphs 257 to 267 above, it reveals at the very least market-sharing conduct, in that the Crédit agricole trader refrained from submitting an offer to a client who had contacted him and who had also contacted the Credit Suisse trader. Moreover, that finding cannot be called into question by Crédit agricole’s argument that the discussion initiated by the Credit Suisse trader was ‘probably motivated by the fact that the same client had already previously asked the [Crédit agricole] trader for the price of the bonds’, as is clear from the case-law referred to in paragraph 186 above. As regards, as a second step, the exchanges of information at issue, first, it is important to note that they concern information which, on the one hand, is not accessible to the public and, on the other hand, is commercially sensitive, contrary to what Crédit agricole maintains in an ad hoc and unsubstantiated manner. The fact that that information was not accessible to the public was established to the requisite legal standard by the Commission in recitals 714 to 719 of the contested decision, according to which only some of that information could be partially discernible via traders’ screens or brokers’ platforms, but, in any event, it was so without making it possible to determine the identity of the ultimate counterparties. According to Crédit agricole, traders ‘constantly and thoroughly research prices and sources of liquidity [in order to] know where they can buy and sell bonds and at what price’, which justifies the exchanges at issue. First, as the Commission rightly pointed out in recitals 43 and 663 of the contested decision, the information exchanged between parties relating to prices currently offered or envisaged goes far beyond what was necessary for the purposes of discovering prices with a view to negotiating bilateral transactions and, secondly, if the information exchanged were actually accessible to the public, there would be no point in exchanging it. In that regard, with regard to the discussion of 10 July 2013 referred to by Crédit agricole, it must be noted that that bank does not provide any evidence - other than a mere assertion - to refute the finding that the information exchanged during that discussion - during which the Crédit agricole trader indicated to the Deutsche Bank trader that he was going to bid for a bond at the same price as that which the latter trader had previously communicated to him - was not public. The commercially sensitive nature of the information exchanged was also established to the requisite legal standard by the Commission as regards information relating both to current and future negotiations and to recent negotiations. In that regard, the Commission noted in particular, in recitals 658 and 663 of the contested decision, that the information exchanged concerned the activities of clients active on the market on a given day, their trading interests and preferences as expressed to each of the banks concerned, including the bonds which the client might hold or seek and their volume, whether the client was likely to continue to buy or sell, and the price levels at which the client wished to execute or had concluded recent transactions. As regards information on recent negotiations, the Commission also noted, in recital 716 of the contested decision, that it could be useful for assessing market trends and future commercial opportunities, as knowledge of past results was highly relevant information for traders, both for monitoring purposes and with a view to future contracts, referring to paragraph 127 (not published) of the judgment of 12 July 2019, Sony and Sony Electronics v Commission (T‑762/15, EU:T:2019:515 ). Although, in view of the specific nature of the SSA bond sector, reference to a case which arose in a sector which is completely foreign to it proves to be of little relevance, the fact remains that the Commission made it fully clear in the contested decision, in particular in recital 659, that the information relating to the recent negotiations was of a commercially sensitive nature, contrary to what Crédit agricole maintains. Moreover, that commercially sensitive nature is confirmed both by the Compass Lexecon report of 29 March 2019, entitled ‘The Economic Context of Price Discussions and the Exchange of “Sensitive Commercial Information” between Market Makers in SSA Bonds’ (‘the first Compass Lexecon report’), produced by Credit Suisse and referred to in recital 705 of the contested decision and by Crédit agricole in its written submissions. In that respect, the latter bank expressly states that, ‘in order to be competitive, a trader needs to understand how prices have moved in the past in order to know what he needs to do to win business’. In that regard, Credit Suisse cannot usefully rely on paragraphs 185 to 191 of the judgment of 24 September 2019, HSBC Holdings and Others v Commission (T‑105/17, EU:T:2019:675 ), in which the Court denied the classification as a ‘restriction by object’ of a discussion between two traders active on the Euro interest rate derivatives market concerning trading positions. It is clear from those paragraphs that the reason why the Court refused to recognise an anticompetitive object in a discussion relating to trading positions is not that such a discussion does not, as a matter of principle, have a sufficiently harmful effect on competition to be classified as a ‘restriction by object’, but that, in that case, the Commission had not shown to the requisite legal standard that the discussion had given traders an advantage in terms of information which could enable them to adjust their trading strategies accordingly. In the present case, it is clear from paragraphs 653 to 655 above that, in the contested decision, the Commission provided such evidence. Furthermore, it should be noted that, in recital 665 of the contested decision, the Commission was careful to distinguish between exchanges of information relating to recent negotiations and those ‘less valuable’ relating to more historical information. Secondly, it is not disputed by either Crédit agricole or Credit Suisse that the exchanges of information in question took place between a small number of operators on the SSA bond market. In that regard, Credit Suisse stated in its written submissions that the advantage of the conduct at issue for the banks concerned was that it took place within a ‘closed group’, enabling those banks ‘to quote each other lower spreads than they would to other market makers, anticipating reciprocity from the trusted counterparty’. Therefore, as the Commission rightly pointed out, in particular in recitals 658, 659, 668 to 670, 672, 680 and 740 to 742 of the contested decision, the exchanges of information at issue led to a reduction in market uncertainty in favour of the banks concerned and, as a result, to an asymmetry of information, enabling them to benefit from ‘significantly more information [than] they would otherwise have possessed’. It is settled case-law that exchanges of information which are capable of reducing or removes uncertainty between participants as regards the timing, extent and details of the modifications to be adopted by the undertakings concerned in their conduct on the market must be regarded as presenting a sufficient degree of harmfulness to competition for it to be considered that it is not necessary to assess their effects (see, to that effect, judgment of 12 January 2023, HSBC Holdings and Others v Commission , C‑883/19 P, EU:C:2023:11 , paragraphs 115 and 116 and the case-law cited). In that respect, it is apparent from paragraphs 279 to 286, 209 to 216, 228 to 232, 233 to 235, 244 to 248 and 249 to 256 above that discussions such as those of 8 February 2012, of 19 March, 24 May, 3 June and 25 July 2013 and of 6 August 2014 have such characteristics. Furthermore, the Court of Justice has held that an exchange of information which is capable of eliminating uncertainty between participants as regards the timing, extent and details of the modifications to be adopted by the undertakings concerned in their conduct on the market must be regarded as having a sufficient degree of harmfulness to competition, regardless of the direct effects in the prices paid by end users (see, to that effect, judgment of 12 January 2023, HSBC Holdings and Others v Commission , C‑883/19 P, EU:C:2023:11 , paragraphs 203 and 204). The Commission was therefore right to take the view that, on the SSA bond market, the exchanges of commercially sensitive information between the banks concerned were sufficiently harmful to competition to contribute to the classification of the conduct at issue, as a whole, as a ‘restriction by object’. Such a conclusion is particularly relevant in view of the fact that, in recital 661 of, and footnote 740 to, the contested decision, the Commission pointed out that the SSA bond market was a market in which the management of risk and uncertainty was one of the key parameters of competition, which is not disputed by the applicants, and rightly relied on the fact that the exchanges between competitors, concerning data which are relevant to the determination of prices and which are not public, were all the more sensitive from the point of view of competition when they took place between traders acting as ‘market makers’ (see, to that effect, judgment of 24 September 2019, HSBC Holdings and Others v Commission , T‑105/17, EU:T:2019:675 , paragraph 145). Furthermore, contrary to what Crédit agricole maintains, a finding that exchanges of commercially sensitive information are sufficiently harmful to competition is not conditional on their being ‘systematic’, as is clear from the case-law referred to in paragraph 663 above. Similarly, Crédit agricole’s allegation, alleging, in essence, that the SSA bond market is a market in which there is significant asymmetry of information between market makers, implying that the increase in that pre-existing asymmetry as a result of the exchanges of information at issue would not be sufficiently harmful to competition, cannot succeed. The relative opacity of the SSA bond market, which means that market makers may find it difficult to determine the fair price of a bond, does not mean that there is an asymmetry of information between those market makers, since they are all confronted with that opacity. In any event, assuming that such asymmetry of information exists, Crédit agricole’s argument comes up against the useful effect that should be guaranteed by the concept of ‘restriction by object’ and, more broadly, by Article 101 TFEU, in so far as it leads to the objective seriousness of conduct being relativised even though it takes place on a market that is particularly sensitive to the conduct of undertakings that do not autonomously determine the policy that they intend to follow on the market. Finally, Credit Suisse cannot usefully rely on the judgment of 23 November 2006, Asnef-Equifax and Administración del Estado (C‑238/05, EU:C:2006:734), to deprive the Commission of the possibility of taking account of the exchanges of commercially sensitive information at issue in order to contribute to the classification of the conduct at issue, as a whole, as a ‘restriction by object’. As has already been pointed out in paragraph 183 above, the exchanges of information which gave rise to the judgment of 23 November 2006, Asnef-Equifax and Administración del Estado (C‑238/05, EU:C:2006:734), were very different from those found and penalised in the contested decision, in particular in that they concerned information which was anonymised and accessible in a non-discriminatory manner to all operators active on the market. In the light of the foregoing, Crédit agricole and Credit Suisse wrongly criticise the Commission for having incorrectly assessed the harmfulness of the conduct at issue in order to conclude that it was capable of contributing to the classification of the conduct at issue, as a whole, as a ‘restriction by object’. (iii) The criticism alleging that the conduct at issue is, in essence, ‘justified’ It follows from the foregoing that the conduct at issue, taken as a whole, is sufficiently harmful to competition to enable the Commission to classify it as a ‘restriction by object’. However, it is necessary to examine whether that finding is likely to be called into question by Crédit agricole’s and Credit Suisse’s arguments based on the ‘justified’ nature of the conduct at issue. In that respect, it should be noted that Crédit agricole argued, indiscriminately, that the discussions of which it was accused were ‘necessary’, ‘legitimate’, ‘pro-competitive’ or even ‘inherent and essential to the proper functioning of the market and the broader servicing of counterparties and clients’. For its part, in support of its claim that the economic context of the infringement at issue had been wrongly assessed, Credit Suisse argued, in an indiscriminate manner, that the discussions at issue were ‘legitimate’ or ‘pro-competitive’. In view of the ambiguous wording used by those two banks, the General Court asked them, as part of a measure of organisation of procedure, to clarify whether their arguments were based on paragraph 103 of the judgment of 30 January 2020, Generics (UK) and Others (C‑307/18, EU:C:2020:52 ), and, if not, to specify the conceptual or case-law basis of their arguments. In its reply, Crédit agricole stated that its ‘main argument is not based on [that] paragraph [103]’ in that ‘the Commission’s failure goes beyond the failure to take account of pro-competitive effects’. However, ‘the conduct of its trader … had beneficial effects, including pro-competitive effects, which the Commission should have taken into account’. In its reply, Credit Suisse stated that its argument, based on the fact that the discussions at issue did not present a sufficient degree of harmfulness to competition to be classified as a ‘restriction by object’, was ‘corroborated by the Court [of Justice]’s considerations in [that] paragraph 103 …, without, however, being based on it’. Nevertheless, at the hearing, in response to a question from the General Court, Credit Suisse indicated that it was indeed relying on paragraph 103. In that context, Crédit agricole’s and Credit Suisse’s arguments must be understood as based essentially on the pro-competitive effects of the conduct at issue and, secondarily, on the ‘necessary’, ‘legitimate’, ‘beneficial’ or ‘inherent and essential’ nature of that conduct. – Crédit agricole’s and Credit Suisse’s criticisms, alleging pro-competitive effects of the conduct at issue As part of its criticism of the pro-competitive effects of the conduct at issue, Credit Suisse claims that the privileged relationships between the traders of the banks which took part in the infringement at issue enabled them not only to improve the quality of the information exchanged between them but also, for themselves and for their clients, to obtain liquidity at better prices. In support of its argument, Credit Suisse relies on the first Compass Lexecon report, according to which price discovery increases competition between market makers and traders by encouraging them to display better prices in order to attract order flow and, subsequently, to reduce transaction costs and spreads. It follows that the Commission could not validly have considered, in recitals 155, 376, 381, 603, 619, 658, 659 and 660 of the contested decision, that the price discovery enabled the banks concerned to assess the competitiveness of their own prices or to assess their correct interpretation of the market as a result of the increased transparency of the market and the reduction of uncertainty on it. Nor could it have validly considered that that price discovery had enabled them, to their advantage, to adjust and determine their strategy with regard to the market and other traders, or to create market asymmetry. Consequently, according to Credit Suisse, the Commission could not conclude that the price discovery was inherently and substantially anticompetitive, or, ultimately, that the conduct at issue fell within the classification of a ‘restriction by object’, particularly since the banks concerned had small combined market shares and the SSA bond market was highly competitive. Crédit agricole also argued that the conduct at issue had a pro-competitive effect. In that respect, it confined itself to stating, in point 2.16 of its application, that ‘the Commission … failed to recognise that the type of information being conveyed between the traders had a range of benefits in terms of facilitating counterparty trades, providing liquidity, and enabling better (and more competitive) pricing, notably to the benefit of clients’. As regards consideration of the alleged pro-competitive effects of conduct, it is apparent from the judgments of 30 June 1966, LTM (56/65, EU:C:1966:38 , p. 249), and of 26 November 2015, Maxima Latvija (C‑345/14, EU:C:2015:784 , paragraphs 16 and 17), that, in assessing the economic and legal context in which the conduct at issue takes place, it is not necessary to examine nor, a fortiori, to prove the effects of that conduct on competition, be they actual or potential, negative or beneficial (judgment of 21 December 2023, European Superleague Company , C‑333/21, EU:C:2023:1011 , paragraph 166). It follows that the pro-competitive effects alleged by the applicants should not, as such, be taken into consideration when classifying the conduct at issue as a ‘restriction by object’. In any event, assuming that the alleged ‘favourable’ implications of that conduct can or must be taken into account, in one way or another, for the purposes of classifying it as a ‘restriction by object’, it is important to note the following. By confining itself to putting forward the argument set out in paragraph 686 above without providing any further details, Crédit agricole cannot prevent the conduct at issue from being classified as a ‘restriction by object’. Similarly, Credit Suisse cannot validly argue, first, that it could limit itself to establishing the ‘prima facie plausibility’ of the ‘favourable’ implications of its conduct and, secondly, that the Commission infringed its rights of defence by applying inappropriate rules of evidence. As regards the importance of the ‘favourable’ implications alleged by Credit Suisse, it should be noted that that bank relies on the ‘favourable’ implications attached solely or essentially to the 25 discussions relating to the determination of prices, which, as has already been pointed out, in paragraphs 110 and 569 above, were unsubstantiated and therefore inadmissible. It is therefore necessary to examine whether, assuming that they have been established, the ‘favourable’ implications of such discussions are such as to preclude the classification as a ‘restriction by object’ of all of the conduct at issue, which the contested decision shows to take the form of exchanges of commercially sensitive information but also of coordination of prices and coordination of trading activity (recitals 614 and 616 of the contested decision). In that regard, Credit Suisse’s argument, which is based essentially on the first Compass Lexecon report submitted by that bank in the course of the administrative procedure, cannot be accepted. First, it should be noted that, according to paragraph 96 of the judgment of 14 September 2022, Google and Alphabet v Commission (Google Android) (T‑604/18, under appeal, EU:T:2022:541 ), the evidential value of statements or reports submitted at the request of a party in support of its allegations by a third party claiming to be an expert is assessed in several respects. On the one hand, the author of the statement or report must set out his or her qualifications and experience and explain how those are relevant to giving an opinion on the matter under review. On the other hand, the content of that opinion must set out the reasons why it deserves to be taken into account, whether it be the reliability of the methodology used or the relevance of the answer given to that question for the purposes of the present case. It is in the light of those principles and the submissions made by the parties in that regard that the Court has considered those documents in the present case. With regard to the second of those conditions, it must be noted that, despite the Court’s invitation to Credit Suisse to indicate the reasons why the first Compass Lexecon report deserved to be taken into consideration, the bank did not provide any details. In addition, it may be noted that the description of the SSA bond sector in points 3.4 to 3.24 of that report is accompanied by virtually no reference to documents capable of substantiating that description. Similarly, with regard to the description of the role of market makers, points 3.32 and 3.33 of that report refer to the central role of market makers, using the corporate bond market as an example and referring to a Commission text that expressly stresses the difference between the corporate bond market and the SSA bond market. Secondly, the conclusions reached in the first Compass Lexecon report are not capable of demonstrating the proven ‘favourable’ implications of price discovery and, a fortiori, the proven ‘favourable’ implications of all of the conduct at issue, which cannot be reduced to such communications. That report does not in any way conclude that price discovery presented proven ‘favourable’ implications. On the contrary, it is clear from points 4.1 and 4.4 of that report that, while noting that such communications are capable of leading to price coordination, the report merely states that, ‘by allowing market makers to price the bond more effectively, discussions about prices and the exchange of “commercially sensitive information” between market makers may be neutral from the point of view of competition or even pro-competitive’. Furthermore, points 4.27 and 4.47 of the first Compass Lexecon report and point A.39 of that report make the possible neutral or ‘favourable’ implications of communications relating to price determination conditional on ‘market participants [having] comparable access to information about real market conditions’ or ‘there being a sufficient proportion of informed operators on the market’. In essence, the first Compass Lexecon report refers to the possible neutral or ‘favourable’ implications of price discovery (as distinct from those aimed at coordinating conduct), only when such communications make the market transparent or more transparent for, at the very least, a sufficient number of operators on the SSA bond market. In the present case, as the Commission rightly pointed out in recital 742 of the contested decision, the conduct at issue created a significant asymmetry of information between, on the one hand, the four banks involved in the infringement at issue and, on the other hand, the other banks operating on the SSA bond market. It found that the asymmetry was created by the fact that each of the banks involved in the infringement at issue was aware of the prices that the other banks concerned would post, as well as the trade flows and specific preferences of clients, of which their competing banks, not involved in the infringement at issue, were completely unaware. That finding, implying the absence of any proven ‘favourable’ implications of price discovery, is ultimately also confirmed by Crédit agricole and Credit Suisse themselves. By relying on the fact that the banks involved in the infringement at issue represented only a small proportion of the market makers active on the SSA bond market, they implicitly admit that the conditions required by the first Compass Lexecon report for communications relating to price determination to be able to generate neutral or ‘favourable’ implications are not satisfied in the present case. Since the ‘favourable’ implications of those communications have not been proven, the same applies to the ‘favourable’ implications of the other conduct at issue, which Crédit agricole and Credit Suisse have not argued in a substantiated manner and, consequently, have not shown to have such implications. In those circumstances, it must be held that Credit Suisse - and a fortiori Crédit agricole - have not shown that the conduct at issue taken as a whole - any more than the conduct relating solely to ‘price discovery’ - had ‘favourable’ implications such as to call into question the classification as a ‘restriction by object’. In recital 713 of the contested decision, therefore, the Commission rightly rejected the argument alleging the pro-competitive or beneficial effects of the conduct at issue. – Crédit agricole’s and Credit Suisse’s criticisms of the ‘necessary’, ‘legitimate’, ‘beneficial’ or ‘inherent and essential’ nature of the conduct at issue As indicated in paragraph 678 above, Crédit agricole and Credit Suisse were asked, as part of the measures of organisation of procedure, to specify the conceptual or jurisprudential basis for their allegations relating in particular to the ‘necessary’, ‘legitimate’, ‘beneficial’ or ‘inherent and essential’ nature of the conduct at issue. Neither the responses to those measures of organisation of procedure, nor the hearings, enabled the Court to determine with any precision the exact basis of those criticisms. It is in that context that the arguments put forward by Crédit agricole and Credit Suisse should be examined, in the event that they make it possible to call into question the classification of the conduct at issue as a ‘restriction by object’. In so far as Crédit agricole and Credit Suisse rely on the ‘necessity’ of the conduct at issue and, therefore, presumably, on the exception for ancillary restraints to a principal transaction, it should be noted that, if a particular transaction or activity does not fall within the prohibition principle laid down in Article 101(1) TFEU, by reason of its neutrality or its beneficial effect on competition, a restriction on the commercial autonomy of one or more of the participants in that transaction or activity does not fall within that principle of prohibition either, if that restriction is an objective necessity for the implementation of that transaction or activity and proportionate to the objectives of either (see judgment of 11 September 2014, MasterCard and Others v Commission , C‑382/12 P, EU:C:2014:2201 , paragraph 89 and the case-law cited). Thus, it must be ascertained first and foremost whether it would be impossible to carry out the activity in question in the absence of the restriction in question (judgment of 11 September 2014, MasterCard and Others v Commission , C‑382/12 P, EU:C:2014:2201 , paragraph 91). In the present case, Crédit agricole and Credit Suisse have in no way demonstrated that SSA bond trading activity, or even SSA bond market-making activity, would have been impossible in the absence of the conduct at issue (see, by analogy, judgment of 24 September 2019, HSBC Holdings and Others v Commission , T‑105/17, EU:T:2019:675 , paragraph 155). The mere fact alleged by Credit Suisse that, owing to the illiquid and opaque nature of the SSA bond market, the provision of immediacy services requires a certain degree of cooperation and regular flows of commercially sensitive information between market makers cannot call that finding into question. On the contrary, first, Credit Suisse itself admits that, in particular, the exchange of information on market makers’ inventory positions makes it ‘easier’ to find opportunities for transactions between traders, which tends to show that, at least for that type of conduct, the exchange of information was not an objective necessity for the banks to carry on their business in the SSA bond sector. In that regard, it is also important to recall that the fact that an activity is merely made more difficult or even less profitable without the restriction concerned cannot be deemed to give that restriction the ‘objective necessity’ required in order for it to be classified as ancillary (judgment of 11 September 2014, MasterCard and Others v Commission , C‑382/12 P, EU:C:2014:2201 , paragraph 91). Secondly, as has already been pointed out in paragraphs 703 to 705 above, the conduct at issue was engaged in by a limited number of banks active in the SSA bond sector, from which it can be inferred with sufficient certainty that all the banks active in that sector were able to carry on their business without resorting to such conduct. In so far as Crédit agricole and Credit Suisse plead that their conduct is ‘legitimate’, ‘beneficial’ or ‘inherent and essential to the proper functioning of the market and the broader servicing of counterparties and clients’, it should be noted that, even if that argument is to be distinguished from that based on the ‘favourable’ or necessary implications of all or part of the conduct at issue, it appears to be capable of being understood as referring to the solution arrived at by the Court of Justice in its judgment of 19 February 2002, Wouters and Others (C‑309/99, EU:C:2002:98, paragraph 97). Crédit agricole does not provide any evidence to show that the reasoning developed in that judgment should be transposed to the present case (see, by analogy, judgment of 27 February 2020, Commission v Belgium (Accountants) , C‑384/18, EU:C:2020:124 , paragraph 49). In that regard, the fact, referred to by Credit Suisse and rejected by the Commission in recitals 707 to 713 of the contested decision, that, on the illiquid and opaque SSA bond market, the banks concerned were systematically at an informational disadvantage compared with counterparties which did not have a permanent presence on the market (smart money clients) and that, as a result, they had to compensate for that lack of information by seeking information from a number of sources, in particular from traders at other banks, is unlikely to preclude classification as a ‘restriction by object’, even if it is proven. As is apparent from paragraphs 124, 127 and 128 of the judgment of 25 March 2021, Lundbeck v Commission (C‑591/16 P, EU:C:2021:243 ), and was pointed out, in essence, in recital 737 of the contested decision, it cannot be accepted that undertakings try to mitigate 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 concerted practices the object of which is to correct those disadvantages. Such factual situations cannot justify an infringement of Article 101 TFEU, still less conduct that has been found to be sufficiently harmful to competition to be classified as a ‘restriction by object’. That must be all the more so in the present case because, first, the conduct at issue was to the detriment of all types of client of the banks concerned and not solely to the detriment of ‘smart money’ clients, as is apparent from recital 710 of the contested decision, and is not disputed by Credit Suisse. Secondly, as the Commission points out in recital 709 of the contested decision, without Credit Suisse disputing it, the external factors influencing the price of SSA bonds are political and economic events affecting the credit risks relating to a country, region or sector, as well as the likelihood of a new bond issue by the same issuer. Furthermore, there is no reason why, in today’s age of instant information and electronic communications, any specific group of investors should have better access to that information. Finally, the vast majority of investors manage a range of assets including equities, bonds and currencies, whereas SSA bond traders specialise in one class of security and are expected to be well informed about relevant external events. Thirdly, the banks concerned were not acting on the SSA bond market solely as market makers and it was clear that their role as traders did not consist solely of ‘promot[ing] liquidity to support the bank’s clients wishing to trade SSA bonds’, as the Commission rightly held in recitals 690 to 697 of the contested decision. That ground, taken together with the fact, noted in recital 19 of the contested decision, that the SSA bond market had no officially designated market makers and therefore that the banks concerned carried on that activity voluntarily without being subject to the obligations to quote two-way prices and to negotiate with counterparties, also justifies the fact that the banks concerned cannot rely on their role as market maker and the risk associated with it to prevent the classification as a ‘restriction by object’, as the Commission rightly held in recitals 712 and 713 of the contested decision. Moreover, the case-law referred to in paragraph 719 above does not apply to conduct which, far from merely having the inherent ‘effect’ of restricting, or at least of potentially restricting, competition by limiting the freedom of action of certain undertakings, has a degree of harmfulness in relation to that competition which justifies considering that its very ‘object’ is to prevent, restrict or distort it. Accordingly, it is only if it transpires, following an examination of the conduct at issue in a given case, that that conduct does not have as its object to prevent, restrict or distort competition, that it must then be determined whether that conduct may fall within the scope of that case-law (see, to that effect, judgments of 28 February 2013, Ordem dos Técnicos Oficiais de Contas , C‑1/12, EU:C:2013:127, paragraph 69; of 4 September 2014, API and Others , C‑184/13 to C‑187/13, C‑194/13, C‑195/13 and C‑208/13, EU:C:2014:2147 , paragraph 49; of 23 November 2017, CHEZ Elektro Bulgaria and FrontEx International , C‑427/16 and C‑428/16, EU:C:2017:890, paragraphs 51, 53, 56 and 57; and of 21 December 2023, European Superleague Company , C‑333/21, EU:C:2023:1011 , paragraph 186). That finding cannot be called into question by the reference made by Credit Suisse, without further explanation, to the judgment of 31 March 1993, Ahlström Osakeyhtiö and Others v Commission (C‑89/85, C‑104/85, C‑114/85, C‑116/85, C‑117/85 and C‑125/85 to C‑129/85, EU:C:1993:120 ), or to paragraph 137 of the judgment of 12 December 2014, Eni v Commission (T‑558/08, EU:T:2014:1080). Accordingly, Credit Suisse cannot usefully rely on the fact that the conduct at issue was intended to enable the four banks to which the contested decision was addressed and to their clients to obtain liquidity at a better price. In that regard, it should also be borne in mind that Article 101 TFEU is intended, like the other competition rules laid down in the Treaty, to protect not only the direct interests of consumers, but also those of competitors and the structure of the market and, in so doing, competition as such (see, by analogy, judgment of 19 March 2015, Dole Food and Dole Fresh Fruit Europe v Commission , C‑286/13 P, EU:C:2015:184 , paragraph 125 and the case-law cited). In the present case, as the Commission pointed out in essence in recital 742 of the contested decision, there can be no doubt that, while the four banks to which the contested decision was addressed and, through them, their clients, had a clear interest in the conduct at issue, that conduct was detrimental not only to the competitors of the banks concerned on the SSA bond market and to their clients, but also to competition as such. In recitals 704 to 713 and 735 to 737 of the contested decision, therefore, the Commission rightly rejected Crédit agricole’s and Credit Suisse’s arguments that the conduct at issue was ‘necessary’, ‘legitimate’, ‘beneficial’ or ‘inherent and essential’ in order to prevent it from being classified as a ‘restriction by object’. (4) Conclusion on the classification of the conduct at issue as a ‘restriction by object’ Having assessed the conduct at issue, taking due account of its content, the objectives it pursued and the economic context in which it took place, the Commission was right, in recital 749 of the contested decision, to find that the agreements and/or concerted practices described in Section 4.2 of the contested decision had to be classified as a ‘restriction by object’ and that it was therefore not for the Commission to demonstrate their anticompetitive effects, in particular those relating to the degree of uncertainty existing on the SSA bond market and resulting from the exchanges of commercially sensitive information in question. Crédit agricole’s first plea in law and Credit Suisse’s first plea in law must therefore be rejected as partly inadmissible and partly unfounded. 4. Crédit agricole’s fourth plea in law and Credit Suisse’s third plea in law, alleging inadequate statement of reasons and errors in determining the amount of their respective fines In determining the amount of the fines imposed on each of the banks concerned, the Commission noted, as a first step, that those banks had infringed Article 101 TFEU intentionally or at least negligently, which entitled it to impose a fine on them pursuant to Article 23(2) of Regulation No 1/2003 (recitals 874 to 886 of the contested decision). As a second step, in the context of the implementation of the Guidelines, and more specifically of the determination of the value of sales which serves as the starting point for the calculation of the basic amount of the fines (points 13 to 18 of the Guidelines), the Commission found, without being contradicted on that point by the applicants, that financial products such as SSA bonds did not generate sales in the usual sense of the term and that, consequently, it was appropriate in the present case to calculate a proxy for the value of sales (recitals 889 and 890 of the contested decision). As a third step, the Commission calculated the basic amount of the fine for each of the banks concerned, first, by taking a proportion of 16% of their respective proxy in respect of the gravity of the infringement at issue (‘the gravity multiplier’) (points 1 to 23 of the Guidelines) (Section 8.2.4.1; recitals 939 to 949 of the contested decision), then by multiplying it by a duration coefficient specific to each of them (‘the duration multiplier’) (point 24 of the Guidelines) (Section 8.2.4.2; recitals 950 and 951), and, finally, by adding an additional amount of 16% for deterrence (point 24 of the Guidelines) (Section 8.2.4.3; recitals 952 to 954). As a fourth step, the Commission adjusted the basic amount calculated for each of the banks concerned (points 27 to 35 of the Guidelines). It did not find any aggravating or mitigating circumstances in respect of any of the banks. By contrast, it multiplied the basic amount of BofA and Crédit agricole by a multiplier for specific deterrence of 1.3 and 1.2 respectively (‘the multiplier for specific deterrence’) (points 30 and 31 of the Guidelines) (recitals 955 to 966 of the contested decision). In the context of the first, third, fourth and fifth parts of Crédit agricole’s fourth plea in law and of Credit Suisse’s third plea in law, consisting of four parts, the applicants challenge the basic amount of the fines imposed on them, essentially on five grounds. First, in the first part of its third plea in law, Credit Suisse complains that the Commission infringed its duty to state reasons as regards the data used to calculate the proxy. Secondly, in the third part of Crédit agricole’s fourth plea in law and in the second part of Credit Suisse’s third plea in law, the applicants claim that the Commission infringed the Guidelines as regards the methods for determining the proxy. Thirdly, in the first part of its fourth plea in law, Crédit agricole complains that the Commission infringed point 13 of the Guidelines, the principle of sound administration and its duty to state reasons, by choosing not to use 2014 - the last full year of that bank’s participation in the infringement at issue - as the reference period for calculating the proxy. Fourthly, in the context of the fourth part of Crédit agricole’s fourth plea in law and of the third part of Credit Suisse’s third plea in law, the applicants criticise the Commission for having committed errors of assessment in determining the gravity multiplier. Fifthly, in the context of the fifth part of Crédit agricole’s fourth plea in law and of the fourth part of Credit Suisse’s third plea in law, the applicants criticise the Commission for having committed errors of assessment in determining the duration multiplier. Moreover, in the context of the second and fourth parts of its fourth plea in law, Crédit agricole challenges the adjustments made by the Commission to the basic amount of the fine imposed on it. More specifically, it complains that the Commission infringed the principle of equal treatment in determining the multiplier for specific deterrence and that it committed errors of assessment in examining the mitigating circumstances from which it should have benefited. Those criticisms will be examined successively and in the order set out in paragraphs 739 to 746 above, which follows the methodology set out in the Guidelines. As a first step, the determination of the basic amount of the fine and, as a second step, the adjustments to that amount will be examined. (a) Criticism of the basic amount of the fines imposed on Crédit agricole and Credit Suisse (1) The criticisms relating to the amount of the proxy As pointed out in paragraph 736 above, the Commission considered that, for the purposes of determining the amount of the fine for each of the banks concerned, it was appropriate to calculate a proxy. In view of the specific nature of SSA bonds, the Commission indicated, in recital 892 of the contested decision, that the starting point for calculating the proxy should be the notional volume and value (‘the notional amounts’) of the SSA bonds which the banks concerned exchanged during their individual period of participation in the infringement at issue. To that end, the Commission used the annualised notional amounts of the SSA bonds negotiated by each bank concerned with counterparties located in the EEA during its period of participation in the infringement at issue, using the following method: ‘the notional amounts traded during the period of individual involvement in the cartel are annualised by a factor [calculated] in function of the duration of the individual participation, dividing the total amount by the number of full months of participation and subsequently multiplying this monthly average by 12’. It therefore decided not to take into account the notional amounts for the last full year of participation by each bank concerned in the infringement at issue, and departed from point 13 of the Guidelines (recitals 893 and 894 of the contested decision). In recital 896 of the contested decision, the Commission then noted that SSA bonds were traded on the secondary market at a price expressed as a percentage of their notional amount. It deduced that the income from those transactions was reflected in the spread between the purchase price and the sale price of each SSA bond acquired and then resold by traders (bid-ask spread) and that it was appropriate to take that spread into account for the purposes of calculating the proxy. To that end, the Commission took the view that it was appropriate to multiply the annualised notional amounts retained for each bank concerned by a factor based on that bid-ask spread (‘the adjustment factor’), namely an average bid-ask spread specific to each bank and reduced by half to take account of the fact that each buy or sale transaction carried out by a bank represented only half of an entire SSA bond buy-sell back transaction (recitals 897 to 899 of the contested decision). To construct the adjustment factor specific to each bank, the Commission first chose to use 33 representative categories of SSA bond issued by eight issuers and divided into five maturity bands (from 0 to 3 years, from 3 to 5 years, from 5 to 7 years, from 7 to 10 years and over 10 years), taking into account the comments made by the banks concerned following the letter on fines (recitals 902 and 931 of the contested decision). Secondly, for each bank and for each day of its participation in the infringement, the Commission reconstructed a daily bid-ask spread, using the method indicated in recital 901 of the contested decision, as follows: For each bank concerned, that daily spread consisted of an average of the daily bid-ask spread for each of the 33 SSA bond categories, weighted according to the notional amounts traded by that bank, by issuer and by maturity. In order to determine the bid-ask spread for each representative category of SSA bond, the Commission used public data from the Bloomberg platform, which it considered appropriate and verifiable (see recitals 903, 926 and 927 of the contested decision). That platform makes it possible to find out, for each of the representative SSA bond categories, the average market price spread at the end of each day (end-of-day spread), calculated from the ‘Bloomberg BGN composite price source’ (‘the BGN data’), based on executable and indicative quotes from several brokers (see recital 903 of, and footnote 1 060 to, the contested decision). Thirdly, the Commission calculated, for each bank concerned, the final bid-ask spread by taking the simple average of the daily bid-ask spread, corresponding to the days of its participation in the infringement at issue (804 days or 574 working days for Crédit agricole and 1 738 days or 1 242 working days for Credit Suisse), in accordance with the method set out in recital 900 of the contested decision, as follows: In recitals 904 and 905 of the contested decision, the Commission thus adopted, for Crédit agricole, a final bid-ask spread of [0.120% - 0.140%] and therefore an adjustment factor of [0.060% - 0.070%] (amount reduced by half), rounded down to the third decimal place, corresponding to a proxy of EUR 6 501 321. For Credit Suisse, it used a final bid-ask spread of [0.190% - 0.210%] and therefore an adjustment factor of [0.095% - 0.105%] (amount reduced by half), rounded down to the third decimal place, corresponding to a proxy of EUR 12 890 852. The details of those calculations are set out in a confidential annex to the contested decision, specific to each bank concerned and notified only to that bank. It is in that context that, first, Credit Suisse complains that the Commission infringed its duty to state reasons, by not providing it with the information enabling it to assess whether the method of calculating the fine was flawed and, secondly, Crédit agricole and Credit Suisse complain, in essence, that the Commission infringed the Guidelines on the grounds that the data used by the Commission did not constitute the ‘best available figures’ within the meaning of point 15 of the Guidelines and therefore did not reflect the economic significance of the infringement at issue. (i) Credit Suisse’s criticism alleging that no reasons were given for the calculation of the fine In the first part of its third plea in law, Credit Suisse complains that the Commission infringed its obligation to state reasons, in that ‘the contested decision does not provide [it] with sufficient information to reproduce the calculation of the adjustment factor’. The Commission failed, first, to send it the BGN data relating to the representative SSA bonds selected and, secondly, to update the International Securities Identification Number (ISIN) references of those SSA bonds, which were changed after the letter on fines was sent. Furthermore, Credit Suisse claims that it attempted to reproduce the Commission’s calculations and determined an adjustment factor not of [0.095% - 0.105%], as retained by the Commission, but of [0.095% - 0.105%] (a difference of one thousandth of a percentage point less). In essence, by the present first part, Credit Suisse complains that the Commission failed to provide the data necessary to enable it to verify whether the contested decision - and more particularly the part of it relating to the calculation of the adjustment factor - was vitiated by a defect such as to compromise the validity of the calculation of the fine imposed on it. As regards, in the first place, Credit Suisse’s allegation that the updated ISIN references were not mentioned in the contested decision, it should be noted at the outset that the Commission cannot validly maintain that ‘the case-law - [namely the] judgment of 15 July 2021, Commission v Landesbank Baden-Württemberg and SRB (C‑584/20 P and C‑621/20 P, EU:C:2021:601 , paragraph 105) - does not require the contested decision to contain information enabling Credit Suisse to reproduce [its] calculations’. In addition to the fact that paragraph 105 of the judgment of 15 July 2021, Commission v Landesbank Baden-Württemberg and SRB (C‑584/20 P and C‑621/20 P, EU:C:2021:601 ), is more nuanced than the Commission tends to suggest, the Court of Justice has repeatedly held that, although that institution is not required to indicate all the figures relating to each of the intermediate stages in the method of calculating the fine imposed on the basis of Article 101 TFEU, it is nevertheless incumbent on it to explain the weighting and assessment which it has made of the factors taken into consideration (see judgment of 10 July 2019, Commission v Icap and Others , C‑39/18 P, EU:C:2019:584 , paragraph 31 and the case-law cited). To that effect, in a context similar to that of the present case, the General Court held that it was essential for the statement of reasons for the contested decision to enable the applicant to verify whether the proxy chosen by the Commission may be vitiated by an error enabling its validity to be challenged (judgment of 24 September 2019, HSBC Holdings and Others v Commission , T‑105/17, EU:T:2019:675 , paragraph 346). However, it must be noted that, in the contested decision and more particularly in the annex thereto, the Commission allowed Credit Suisse to identify precisely the representative SSA bonds taken into consideration for the calculation of the adjustment factor, even though it did not refer to the updated ISIN references of those SSA bonds. It is clear from the annex to the contested decision, entitled ‘Annex to Decision AT.40346 SSA bonds (Credit Suisse) (confidential data)’, that the Commission enabled Credit Suisse to verify the correctness of its calculations by indicating, in addition to the annualised notional amount of that bank’s transactions (Table 1), the components of the adjustment factor, namely the weighting by issuer (Table 2), the weighting by maturity (Table 3) and the weight matrix (Table 4). In addition, as regards the identification of the representative bonds used to calculate Credit Suisse’s adjustment factor, the Commission provided a table for each of the 33 categories of SSA bond concerned (Tables 5 to 12), presented in the following form: For each of those 33 categories of SSA bond, the Commission identified several specific SSA bonds covering the different periods of the infringement at issue, which led it to identify 88 specific SSA bonds, as Credit Suisse pointed out. As the Commission rightly pointed out, such a presentation makes it possible to identify, with sufficient precision, the issuer of the SSA bond concerned, the coupon rate and the maturity date, for each period of the infringement at issue identified with its start date and end date, which Credit Suisse also expressly admitted in paragraph 82 of its reply. Accordingly, the Commission cannot validly be criticised for having failed in its duty to state reasons by not mentioning or updating, in the contested decision, the ISIN references of the representative SSA bonds used to calculate the adjustment factor. Furthermore, it may be noted that the Compass Lexecon report of 7 July 2021, attached to the Credit Suisse application and entitled ‘Economic report on the calculation of Credit Suisse’s bid-ask spread adjustement factor’, shows that, by means of the contested decision and its annex or of the letter on fines, that bank was able to identify with certainty the ISIN reference of 85 of the 88 representative SSA bonds used to calculate the adjustment factor. With regard to the three remaining representative SSA bonds, Credit Suisse states in the same report that each of them was likely to receive two ISIN identifiers. It notes, however, that ‘for each of them, there was only one ISIN reference that corresponded to a bond for which bid and ask price data were available on BGN’. It follows that Credit Suisse had no difficulty in understanding that the Commission, which had indicated that it had extracted the data used to calculate the adjustment factor from the BGN data, had used the SSA bonds for which the data were available. In addition, the Compass Lexecon report of 11 February 2022, entitled ‘The inability to replicate the calculation of Credit Suisse’s bid-ask spread adjustement factor’, states that ‘regardless of the combination of ISINs below used for these Representative Bonds, the overall average bid-ask spread that we calculate is always [the same]’. Consequently, despite any inaccuracies in the contested decision and its annex or in the letter on fines which Credit Suisse itself admits do not alter the result of its calculation, that bank was able to identify all the representative SSA bonds used by the Commission to calculate the adjustment factor and to verify their accuracy and validity. As regards, in the second place, Credit Suisse’s allegation that the contested decision does not contain the data taken into account in the calculation of the adjustment factor - namely, the BGN data of the representative SSA bonds used to calculate that factor - that fact is not sufficient to establish a failure by the Commission to fulfil its obligation to state reasons for the contested decision. In that regard, it should be borne in mind that it is not required that the statement of reasons for an EU act specify all the relevant matters of fact and of law, in so far as, first, the sufficiency of a statement of reasons must be assessed in the light not only of its wording but also of its context and of all the legal rules governing the matter concerned and, secondly, the fact that the act complained of occurred in a context known to its addressee must be duly taken into account (see, to that effect, judgment of 15 July 2021, Commission v Landesbank Baden-Württemberg and SRB , C‑584/20 P and C‑621/20 P, EU:C:2021:601 , paragraph 104 and the case-law cited). In the present case, it has already been noted in paragraph 775 above that the contested decision identified with sufficient precision the representative SSA bonds taken into account for the calculation of the adjustment factor. Moreover, as is apparent from recital 927 of the contested decision and has not been disputed by Credit Suisse, the BGN data are public data and therefore accessible to Credit Suisse. Consequently, there was no need for the Commission to refer in the contested decision to all of the daily BGN data of all the representative SSA bonds used to calculate the adjustment factor. Furthermore, the fact that Credit Suisse was unable to reproduce the Commission’s calculation of the adjustment factor cannot, in the present case, establish that the contested decision is inadequately reasoned on that point. In its rejoinder, and without being contradicted by Credit Suisse at the hearing, the Commission stated in great detail that the different result obtained by that bank was due to two calculation errors made by Credit Suisse, resulting from the erroneous application of Tables 5 and 8 of the annex to the contested decision, relating to the SSA bond categories issued by the EIB and BNG respectively. Finally, even if, contrary to both the title and the letter of the first part of its third plea in law, Credit Suisse relies on the fact that the Commission did not send it, during the administrative procedure, the BGN data of the representative SSA bonds used to calculate the adjustment factor and their updated ISIN references, such a criticism is irrelevant to the failure to state reasons for the contested decision, which is alleged in the present part. In the light of the foregoing, the first part of Credit Suisse’s third plea in law must be rejected as unfounded. (ii) Crédit agricole’s and Credit Suisse’s criticisms, alleging breaches of the Guidelines In the context of the criticisms levelled by Crédit agricole and by Credit Suisse, alleging breaches of the Guidelines in the determination of the proxy, those two banks, taken together or separately, criticise the Commission, first, for having relied on a set of representative SSA bonds and thus for not having used a methodology for calculating the spread based on data from their own transactions (second argument in the third part of Crédit agricole’s fourth plea in law and second argument in the second part of Credit Suisse’s third plea in law); next, that it used a fixed weighting of the SSA bonds for the whole of the infringement period and for all of the banks concerned for the purposes of calculating the final spread for each of those banks (first argument in the third part of Crédit agricole’s fourth plea in law) and, finally, that it used BGN data, which overestimated the spreads for the SSA bonds (third argument in the third part of Crédit agricole’s fourth plea in law and first argument in the second part of Credit Suisse’s third plea in law). In the alternative, Credit Suisse complains that the Commission attributed to it notional amounts from other banks (third argument in the second part of its third plea in law). Those four criticisms will be examined in turn. – Crédit agricole’s and Credit Suisse’s criticisms, alleging the Commission’s refusal to use a methodology for calculating the spread based on data from their own transactions As a preliminary point, it should be noted, as already noted in paragraphs 748 to 760 above, that in the contested decision the Commission considered it appropriate, for the purposes of determining the basic amount of the fine, to have recourse to a proxy, determined by multiplying the annualised notional amounts of the SSA bonds exchanged by each bank concerned during its period of participation in the infringement at issue, by an adjustment factor, calculated by retaining 50% of a final bid-ask spread specific to each of those banks, calculated by taking the simple average of the weighted daily bid-ask spread of 33 representative categories of SSA bond traded during the period of its participation in that infringement. It is apparent from recital 902 of the contested decision, read in conjunction with points 18 to 22 of the letters on fines addressed to Crédit agricole and Credit Suisse, to which that recital refers, that the Commission selected a sample of 33 categories of SSA bond issued by 8 issuers, among the top 10 issuers by notional amounts traded for at least three banks and divided into five maturity bands. It explains that choice on the grounds of the large number of issuers of SSA bonds on the market, some of which are very small, and for practical reasons, while noting that, for each day of the infringement period, the representative SSA bonds selected are the same for all of the banks and are derived from the breakdown of the notional amounts traded by maturity and by issuer provided by the banks. In recitals 930 and 931 of the contested decision, the Commission also stated that it had adjusted the weighting of issuers and maturities used in its sample compared with that initially used in each letter on fines, in order to take account of the comments made by BofA, Crédit agricole and Credit Suisse on those letters. In that regard, concerning Crédit agricole, the Commission stated that it had taken account of the breakdown by issuer and by maturity of the SSA bonds held by that bank, as communicated by it. That breakdown is presented in Tables 2 and 3 of the annex to the contested decision, which was not provided by the bank as an annex to its application, but by the Commission as an annex to its defence. Those two tables, entitled ‘Weight per issuer (Crédit agricole)’ and ‘Weight per maturity (Crédit agricole)’, read as follows: The data contained in Tables 2 and 3 were cross-referenced by the Commission to establish its matrix for calculating the adjustment factor, set out in Table 4 of the annex to the contested decision, to which the BGN data were applied for each day of Crédit agricole’s participation in the infringement at issue. Furthermore, in the annex to the contested decision, the Commission indicated that it had not included in the sample the SSA bonds issued by the Agence française de développement (AFD), which is present only in Crédit agricole’s top 10 issuers and not in those of the other parties. With regard to Credit Suisse, the Commission provided details similar to those referred to in paragraphs 794 and 795 above, although that bank did not contest them in its action. In the context of the second argument of the third part of its fourth plea in law, Crédit agricole claims that the Commission departed from point 15 of the Guidelines in that it did not use the ‘best available figures’ by having recourse to a sample of 33 categories of SSA bond, which it considers to be representative, but does not explain whether the SSA bonds negotiated by that bank are in fact representative. Thus, according to Crédit agricole, almost 99% of the SSA bonds taken into account by the Commission ‘were not traded by [that bank] at any time during its period of participation in the alleged infringement or were not traded on the day on which they are supposed to be representative’, so that the SSA bonds taken into account in its respect did not reflect its actual trading activity. Consequently, the Commission should have used the SSA bonds actually traded by Crédit agricole during its period of participation in the infringement at issue, especially as it would have had details of them and such an approach would not have required much more calculation. Alternatively, the Commission could simply have selected the bonds most frequently traded by Crédit agricole during its period of participation in the infringement at issue, which would also have involved relatively little additional work and would have produced a much more representative result. Similarly, in the second argument of the second part of its third plea in law, Credit Suisse submits that the Commission should have based the calculation of its final bid-ask spread on data from its own transactions, in accordance with the alternative methodology contained in the Compass Lexecon report of 8 January 2021, entitled ‘Economic response to the EC’s Letter’ (‘the second Compass Lexecon report’), produced by that bank in the course of the administrative procedure and annexed to its application. That methodology, which the Commission rejected in recitals 923 and 924 of the contested decision, reflects the over-the-counter nature of the SSA bond market and the fact that Credit Suisse’s bid-ask spreads are narrower than those derived from BGN data. The alternative methodology proposed by Credit Suisse would be based on data from that bank’s transactions which, during a given period, were the subject of at least one purchase and one sale, namely either ‘perfectly balanced’ trades where the theoretical amounts of the buy and sale transactions were the same, or ‘partially balanced’ trades where those amounts were different. Using those data, Credit Suisse calculated, for each SSA bond and time horizon, an average buy price and an average sell price based on the weighted average of the prices of all fully or partially balanced trades. In addition, Credit Suisse took into account the difference between the average buy and sell prices to estimate the bid-ask spread that it could actually have obtained. Furthermore, according to Credit Suisse, its alternative methodology did not have the weaknesses referred to by the Commission in recitals 923 and 924 of the contested decision. First, the average delay of 107 minutes between the daily purchase and sale of ‘partially balanced’ trades was not sufficient to make its calculation methodology less appropriate than that of the Commission. Secondly, the fact that the ‘partially balanced’ trades do not involve the same amounts would not lead to a distorted measure. The gaps between ‘balanced’ and ‘partially balanced’ trades would be very similar, resulting in little distortion. Thirdly, the Commission’s assertion that ‘balanced trades’ do not constitute a representative sample would be inaccurate, given the similarity of the bid-ask spreads referred to in the previous sentence, and would be irrelevant given that ‘the BGN mid prices are very close to the mid prices derived from Credit Suisse’s actual trade data’. Fourthly, in its previous decision-making practice, the Commission had already used data relating to actual sales, preferring them to publicly available data. By means of all the arguments set out in paragraphs 798 to 804 above, Crédit agricole and Credit Suisse, in essence, criticise the Commission for the principle of using a sample of SSA bonds as a reference for the purposes of calculating the adjustment factor for the fines imposed in the contested decision and, in the case of Crédit agricole, for the unjustified and inappropriate nature of the sample of 33 categories of SSA bonds chosen by that institution. In that regard, it should be noted from the outset that point 15 of the Guidelines requires the Commission to use the ‘best available figures’ in order to determine the value of an undertaking’s sales, as defined in point 13 of those guidelines, and subject to point 16 of those guidelines. Consequently, in the context of the present case, where the Commission has chosen to depart from point 13 of the Guidelines and has therefore refrained from determining the basic amount of the fine using the value of the sales of goods or services made by the undertaking in direct or indirect relation to the infringement, point 15 of those guidelines cannot be directly applicable. Moreover, it is settled case-law that the Commission enjoys a broad discretion as regards the calculation of fines in relation to infringement of the EU competition rules (judgment of 10 July 2019, Commission v Icap and Others , C‑39/18 P, EU:C:2019:584 , paragraph 25). While the Commission has admittedly limited itself in the exercise of that discretion by adopting the Guidelines, it is entitled to depart from them, provided that it gives reasons for doing so (see, to that effect, judgment of 10 July 2019, Commission v Icap and Others , C‑39/18 P, EU:C:2019:584 , paragraph 29 and the case-law cited). In that regard, the Court of Justice has held that the method, described by the Guidelines and based on taking into account the value of the sales of the products concerned in relation to the infringement for the purpose of setting the basic amount of the fines to be imposed, may sometimes prove unsuited to the particular circumstances of a case, and that, in a situation of that kind, the Commission is entitled to have recourse to a calculation method other than that described in the Guidelines (see, to that effect, judgment of 10 July 2019, Commission v Icap and Others , C‑39/18 P, EU:C:2019:584 , paragraph 27). It follows that, provided that it not only gives reasons for its choice but also justifies it to the requisite legal standard, the Commission cannot be deprived of the possibility, for the purposes of determining the basic amount of a fine, of using a method which takes account of samples which are representative of the activities of the undertakings concerned. The fact that the data obtained from those representative samples are not identical to the data obtained from the analysis of their own trades is inherent in the method used and cannot prevent the Commission from using such a calculation method, provided that the samples taken into consideration are indeed representative. In that respect, it should be noted that the case-law accepts that the value of sales referred to in point 13 of the Guidelines neither is nor should be a perfectly accurate reflection of the scale of the infringement at issue, and the same applies to a proxy for the value of sales. While approving the methodology laid down by the Commission in point 13 of the Guidelines, the Court acknowledges, first, that the proportion of turnover derived from the goods or services which are the subject matter of the infringement is only of such a kind as to provide an indication of the scale of the infringement and to reflect the economic importance of that infringement and, secondly, that it must not be given disproportionate importance (see, to that effect, judgment of 19 March 2009, Archer Daniels Midland v Commission , C‑510/06 P, EU:C:2009:166, paragraphs 74 and 76 and the case-law cited). Along the same lines, the General Court has held that setting a fine under Article 101 TFEU is not a precise arithmetical exercise (judgments of 5 October 2011, Romana Tabacchi v Commission , T‑11/06, EU:T:2011:560, paragraph 266, and of 15 July 2015, SLM and Ori Martin v Commission , T‑389/10 and T‑419/10, EU:T:2015:513, paragraph 436). However, where, in the exercise of its broad power to assess the methodology for calculating a fine, the Commission has chosen to depart from the Guidelines not in their entirety - as point 37 thereof authorises it to do - but only, as in the present case, from point 13 thereof, it cannot depart from the guiding principles and the underlying logic of those guidelines (see, to that effect, judgment of 24 September 2019, HSBC Holdings and Others v Commission , T‑105/17, EU:T:2019:675 , paragraphs 324 and 345) and, in particular, the principle expressed in point 15 thereof. Thus, in implementing the methodology which it lays down, it is incumbent on it, inter alia, to ensure that it takes account of the best available figures, subject to detailed review, in law and in fact, by the EU Courts (see, to that effect, judgments of 8 December 2011, Chalkor v Commission , C‑386/10 P, EU:C:2011:815 , paragraph 62, and of 24 September 2019, HSBC Holdings and Others v Commission , T‑105/17, EU:T:2019:675 , paragraph 348). It is in that context that it must be assessed whether the Commission has given sufficient reasons and legal justification for having set aside the methodology provided for in point 13 of the Guidelines and replaced it with a methodology leading it to calculate a proxy 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 described in recital 902 of the contested decision. In that regard, in the first place, it should be noted that, in recitals 887 to 890 of the contested decision, the Commission explained, without that being contested by the applicants, the reasons why it used a proxy as the starting point for determining the basic amount of the fine. As regards the use of a sample of 33 categories of SSA bond issued by the eight issuers referred to in recital 902 of the contested decision, it is clear from paragraphs 792 to 796 above that the Commission has fulfilled its obligation to state reasons, not only as regards the reasons why it selected those eight issuers of SSA bonds and excluded certain other issuers, in particular those in Crédit agricole’s top 10, but also as regards the reasons why it selected the maturity dates of the SSA bonds issued by those eight issuers. As regards, in the second place, the appropriateness of the Commission’s choice of the 33 categories of SSA bond and hence their representativeness, it should be noted that Credit Suisse does not contest them. Crédit agricole, for its part, does not question the representativeness of the sample in relation to all of the banks involved in the infringement at issue, but maintains that the sample is not representative of its own trading activity. First, it is clear from the column ‘Selected issuers’ in Table 2 of the annex to the contested decision, reproduced in paragraph 794 above and not contested by Crédit agricole, that the issuers selected by the Commission correspond to 78.8% of the SSA bond issuers in that bank’s portfolio. Secondly, it should be noted that Tables 3 and 4 of the annex to the contested decision, reproduced in paragraphs 794 and 795 above and which Crédit agricole does not dispute either, show that the calculation matrix used by the Commission led it to apply, to the eight SSA bond issuers retained in respect of all of the banks concerned, a maturity weighting specific to each of those banks and defined on the basis of the data transmitted by those banks during the administrative procedure. Accordingly, the sample of 33 categories of SSA bond selected by the Commission is sufficiently representative, in particular, of Crédit agricole’s activity, especially since, at other stages in the calculation of the fine, the Commission further individualised the proxy used for each bank. The adjustment factor used for each bank is multiplied by the annualised notional amounts specific to each bank, which include all of the SSA bonds that they have exchanged and not just the notional amounts of the 33 categories of representative SSA bonds used by the Commission to calculate the adjustment factor. Moreover, the finding that the 33 categories of SSA bond selected by the Commission are sufficiently representative is not called into question by Crédit agricole’s assertion that nearly 99% of those SSA bonds ‘were not traded by [that bank] at any time during its period of participation in the alleged infringement or were not traded on the day on which they are supposed to be representative’. It should be noted that that assertion is not supported by any evidence, including figures, which would enable the General Court to assess whether it is well founded. If, by that argument, Crédit agricole relies on the fact, which has not been demonstrated, that it did not negotiate or negotiated little with each of the 86 - and not, in its case, 88 - specific SSA bonds taken into account within each of the 33 categories of representative SSA bonds, as presented in paragraphs 771 and 772 above, it should be noted that that bank has not demonstrated how the purchase or sale prices of the various specific SSA bonds taken into account for each of those 33 categories of representative SSA bonds would differ from the buy or sale price of other SSA bonds issued by the same issuer and having the same maturity as those taken into account for each of those 33 categories. Furthermore, in the light of the information available to the Court and, in particular, Tables 2 to 4 in the annex to the contested decision, which were drawn up on the basis of the figures communicated by Crédit agricole itself during the administrative procedure and which that bank does not dispute, the merits of Crédit agricole’s criticism are all the less proven since the representative nature of an SSA bond implies that a bank could have negotiated other SSA bonds without the methodology followed by the Commission being subject to criticism. As regards, in the third place, the alternative calculation methodologies proposed by Crédit agricole and Credit Suisse, it should be noted that those two banks criticise the Commission for not having adopted a methodology for calculating the adjustment factor based on their own transactions and that, in the alternative, Crédit agricole criticises the Commission for not having adopted a methodology based on the SSA bonds most frequently traded by that bank during its period of participation in the infringement at issue. First, a methodology based on the transactions of Crédit agricole and Credit Suisse would lead the Commission, as it rightly pointed out in recital 923 of the contested decision, to calculate an average bid-ask spread using buy-sell or sell-buy transactions which are not comparable, in at least two respects. Points 2.10 and 2.12 of the second Compass Lexecon report show that the alternative methodology proposed by Credit Suisse is based essentially on partially balanced trades, which are defined as trades in bonds which, over a certain period (a day, a week or a month), have been both bought and sold by Credit Suisse traders. On the one hand, Credit Suisse itself admits in its argument concerning the Commission’s use of BGN data that it is important to take account of the volumes traded, which are likely to have an influence on the bid-ask spread between the SSA bonds traded. Consequently, the comparison of buy-sell transactions of different amounts is liable to distort the calculation of the bid-ask spread, a defect which does not affect the Commission’s methodology, pursuant to which it compares amounts aggregated by the Bloomberg platform, as is apparent from footnote 1084 to the contested decision. On the other hand, as the Commission points out, in essence, in recital 923 of the contested decision, given the volatility of SSA bond prices, as admitted by Credit Suisse, the comparison of SSA bond prices which may be separated by ‘a day, a week or a month’ (according to the definition of partially balanced trades given in the second Compass Lexecon report) is likely to distort the calculation of the bid-ask spread. In that respect, the fact pointed out by Credit Suisse that the average time between buy and sell transactions in partially balanced daily trades would be 107 minutes is not sufficient to compensate for that weakness. That is because it does not make it possible to determine the average time for weekly or monthly transactions that are also included in partially balanced trades, or the proportion of partially balanced daily trades compared with partially balanced weekly and monthly trades. The Commission’s methodology does not suffer from that flaw either. It compares amounts relating to the buy and sale prices of SSA bonds, aggregated by the Bloomberg platform and calculated at the same time, namely at the end of the day, as is clear from recital 921 of the contested decision and is not disputed by Credit Suisse. Secondly, the methodology based on their own transactions, as proposed by Crédit agricole and Credit Suisse, would lead the Commission to rely on a sample of transactions whose representativeness has not been demonstrated. It is clear from Crédit agricole’s statements, reported in recital 926 of the contested decision, that the bank is not in a position to have all of the prices of its transactions. In the course of the administrative procedure, it argued that ‘in the absence of actual transaction prices with which to calculate the bid-ask spread, price-spread indices (such as BGN) that are both bond specific and time specific should be used as proxies’. Similarly, point 2.12 of the second Compass Lexecon report shows that, depending on the period, Credit Suisse’s partially balanced trades, taken as a reference for the purposes of that report, accounted for only 14.2% to 59.2% of that bank’s trades, while fully balanced trades accounted for only 1% to 2%. Credit Suisse cannot validly rely on an alternative calculation methodology which is based on data which, at certain times during its participation in the infringement at issue, relate to transactions which are representative of only about 15% of that bank’s total transactions. Nor is Credit Suisse justified in concluding that, where partially balanced trades account for a significant proportion of all trades (‘between 14.2% and 59.2%’ according to point 2.27 of the second Compass Lexecon report), it can assume that the set of trades for which spreads have been calculated is representative of all of the transactions concerned. On that point, on the contrary, the Commission’s methodology does not suffer from that defect in so far as, as is clear from recital 903 of the contested decision, it selected a sample of SSA bonds that were representative of each bank’s portfolio, in terms of both issuers and maturities. Thus, that methodology tends to represent the entire activity of the sanctioned banks and not simply their partially balanced trades. Thirdly, a methodology based on the transaction data of Crédit agricole and Credit Suisse would deprive the Commission of the possibility of maintaining a unity factor in its method of calculating the proxy between all of the banks concerned - that is to say, the sample of 33 categories of representative SSA bonds applicable indiscriminately to all of those banks - even though the basic amount of a fine is intended to reflect not only the economic importance of the infringement at issue, but also the relative weight of each undertaking participating in the infringement at issue, as is clear from point 6 of the Guidelines. Fourthly, a methodology based on the transaction data of the banks concerned would also entail carrying out calculations of a much greater complexity than the already complex ones carried out in the present case, even though the representative nature of the SSA bonds selected, as noted in paragraph 824 above, precisely guarantees that the data taken into consideration remain relevant for the calculation of the fine and make it possible to reflect the economic significance of the infringement at issue with the degree of precision required by the case-law referred to in paragraph 813 above. The methodology based on the transaction data of the banks concerned would mean requiring the Commission to identify and apply individually, for each bank concerned, a yield spread specific to each transaction identified for each day on which that bank participated in the infringement at issue, as is attested by Crédit agricole’s proposal to adopt a ‘methodology … based on [that bank’s] transaction data concerning bonds which, during a given period, were the subject of at least one purchase and one sale’. In addition, it would require the Commission to identify, for each of the banks concerned and for each transaction, corresponding buy and sale transactions that would enable it to calculate a representative price spread, in a context where, as pointed out in paragraphs 833 and 835 above, the amount of the transactions has an impact on their price and where the intraday volatility of the price of SSA bonds makes it difficult, if not impossible, to match transactions taking place at different times. Such a methodology - if it did not make it excessively difficult, if not impossible, to calculate the adjustment factor - would place a disproportionate administrative burden on the Commission (see, by analogy, Opinion of Advocate General Kokott in Pilkington Group and Others v Commission , C‑101/15 P, EU:C:2016:258 , point 37). Fifthly, the fact that, in the case which gave rise to Decision C(2016) 8530 final of 7 December 2016 relating to a proceeding under Article 101 [TFEU] and Article 53 of the EEA Agreement (Case AT.39914 - Euro Interest Rate Derivatives), the Commission was able to use the data of the banks concerned could not require it to use them in the contested decision. Apart from the differences between the two cases, which the Commission noted in recital 924 of the contested decision, it is settled case-law that that institution is not bound by its previous administrative practice (see, to that effect, judgment of 19 March 2009, Archer Daniels Midland v Commission , C‑510/06 P, EU:C:2009:166, paragraph 82 and the case-law cited). With regard to the alternative method proposed by Crédit agricole, it should be noted that it is based on a selection of SSA bonds, which, in Crédit agricole’s own opinion, would infringe point 15 of the Guidelines. In addition, apart from the fact that Crédit agricole does not identify the SSA bonds that it considers to be representative, that method suffers from the same defects as those mentioned in paragraphs 831 to 850 above. Accordingly, by using the sample of 33 representative categories of SSA bond, referred to in recital 902 of the contested decision, in order to determine the proxy serving as a basis for calculating the fines imposed on the banks concerned, the Commission did not exceed the margin of discretion conferred on it for the definition of the method of calculating the fines, in a context in which, as is not disputed by the applicants, the value of sales did not constitute an adequate indicator for the purposes of calculating that basic amount. As the Commission rightly pointed out in recital 933 of the contested decision, the methodology described in recitals 889 to 905 of the contested decision is clear and consistent and allows for a transparent and workable determination of the proxy. It can be applied consistently to all parties, and the level of detail inherent in each element is appropriate for an efficient, credible and understandable calculation of the proxy. Moreover, as the Commission also correctly held in recital 936 of the contested decision, the alternative methodologies proposed by Crédit agricole and Credit Suisse cannot be more appropriate than the one adopted by the Commission in the contested decision. In the light of the foregoing, the second argument in the third part of Crédit agricole’s fourth plea in law and the second argument in the second part of Credit Suisse’s third plea in law must be rejected as unfounded. – Crédit agricole’s criticism, alleging use of fixed SSA bond weightings for the entire infringement period and for all of the banks concerned for the purposes of calculating the final bid-ask spread for each of those banks In the context of the first argument of the third part of its fourth plea in law, as set out in its application, Crédit agricole complains that the Commission ‘calculated the [final] bid-ask spread … by using fixed weights across [all of] the relevant periods for each of the banks, and not just the relevant period as it applies to [Crédit agricole]’, and ‘[therefore] failed to take account of variations in trading patterns, generally and in relation to [Crédit agricole]’s own trading’. Crédit agricole’s criticism, directed, in essence, against recital 933 of the contested decision, stems from a misunderstanding of the methodology adopted and applied by the Commission, as set out in particular in recitals 902 and 903 of the contested decision. As is apparent from recital 903 of the contested decision, it is admittedly true that the Commission used the same sample of 33 representative categories of SSA bond for the purposes of calculating the proxy for all of the banks concerned and that the breakdown of SSA bonds by issuer and by maturity for a given bank remained the same for the entire period of its participation in the infringement at issue. However, it is clear from that recital 903 that, for each party and for each working day of their respective infringement period, the Commission calculated a daily bid-ask spread based on a weighted average of those 33 categories of representative SSA bonds. As regards in particular Crédit agricole, that is also clear from the column ‘Rescaled weights’ in Table 2 of the annex to the contested decision, read in conjunction with Tables 3 and 4 of that annex, reproduced in paragraphs 794 and 795 above. Those tables highlight the fact that, although the reference SSA bonds were the same for all of the banks concerned throughout the infringement period and the breakdown of SSA bonds by issuer and by maturity for a given bank remained the same for the entire period of its participation in the infringement at issue, that breakdown was specific to each bank and gave rise to the determination of a daily bid-ask spread which was therefore also specific to it and which, as a result, took account of the specific features of its trading activity. Moreover, since, in its reply, Crédit agricole states that its argument was intended to criticise the Commission’s failure to take account of the ‘inter-day weighting’, it must be noted that that argument - even supposing that it could be considered admissible despite its late submission - cannot succeed. In its argument, Crédit agricole confines itself, on the basis of what it calls a ‘hypothetical’ example, to stating that, in the event of significant variations in a bank’s trading intensity from one day to the next, the method used by the Commission would tend to overestimate the banks’ activity. That bank has not provided any evidence to show that its activity had exhibited such characteristics during the period in which it participated in the infringement at issue. In the light of the foregoing, the first argument in the third part of Crédit agricole’s fourth plea in law must be rejected as unfounded. – Crédit agricole’s and Credit Suisse’s criticisms of the use of BGN data In recital 903 of the contested decision and in footnote 1060 to which that recital refers, the Commission stated that ‘bid-ask spread data … for [the] 33 [categories of] representative bonds were collected from the Bloomberg BGN composite price source[, described as] a real-time composite price for corporate and government bonds’. In recital 918 of the contested decision, the Commission justified the use of BGN data as follows: ‘As regards the source of information on price spreads, the letter [on fines] explained that …, after analysis: “The data contained on the file does not provide a sample that would be sufficient in itself to ensure that all relevant maturities and issuers declared by the parties in their responses to the Commission’s requests for information are proportionately represented” and that the Commission has therefore collected bid-ask spread data from the Bloomberg BGN composite price source on a daily basis for the representative bonds. In other words, having provisionally foreseen the use of both data on its file and public data and having determined that the file data was not sufficiently representative, the Commission chose the most appropriate data source for calculating the applicable price spreads. …’ In footnote 1084 under recital 921 of the contested decision, the Commission stated that, ‘although for present-day prices, Bloomberg BGN prices are real-time prices and are computed intraday as new quotes arrive …, for historical prices only the end-of-day bid price and ask price are available’. In footnote 1085 under that recital in the contested decision, the Commission added the following: ‘… the Commission used only Bloomberg BGN bid and ask prices to compute the daily bid-ask spreads. Bloomberg BGN bid and ask prices are estimated at the same moment (end of day). Consequently, computing a spread by substracting the BGN bid price from the BGN ask price does not suffer from a time gap that would potentially bias any resulting spread. However, this end day spread cannot be applied to intra-day prices … In contrast, the Commission’s approach is to use the average end day spread for a particular representative bond throughout the period of participation and apply to the notional amount traded.’ In the third argument of the third part of Crédit agricole’s fourth plea in law and in the first argument of the second part of Credit Suisse’s third plea in law, the applicants criticise the Commission for having used the BGN data, which are inadequate for the purposes of calculating the proxy. Those data inflated the price at which the transactions were actually carried out by the banks concerned and, consequently, the adjustment factor calculated on the basis thereof. In that regard, Crédit agricole states, in particular, that on a ‘limited sample’ of trades, the bid-ask spreads actually obtained are about four times lower than the end-of-day bid-ask spreads derived from BGN data. This would constitute proof that those bid-ask spreads were artificially higher than the actual bid-ask spreads. In those circumstances, the Commission should have taken account of those ‘best available figures’, even if they could only be used for indicative purposes, in particular ‘by reducing the final [bid-ask] spread by an appropriate factor’. For its part, Credit Suisse criticises the Commission for having used the BGN data, which are based on an unspecified and opaque ‘proprietary methodology’. In addition, the data appear to reflect initial quotes and price estimates rather than transaction spreads, which can only provide the maximum spreads that the bank can expect to obtain when negotiating with clients, as appears from the second Compass Lexecon report. The BGN data would also be incomplete in that it would not take into account the potential trading volumes that can have an impact on bid-ask spreads. In addition, they would overestimate the spreads obtained when Credit Suisse conducts transactions with other market makers, as would also appear from the second Compass Lexecon report. Furthermore, Credit Suisse takes the view that the Commission cannot disregard the most appropriate source of data on the grounds that another party to the cartel cannot produce the same data. As pointed out in paragraph 807 above, point 15 of the Guidelines is not directly applicable in the present case, in so far as the Commission has departed from the method for calculating fines set out in point 13 of those guidelines. However, where the Commission, in exercising its wide discretion as to the method of calculating a fine, has in the present case rightly defined a method of calculating fines other than that provided for in point 13 of the Guidelines, it must nevertheless ensure that it takes account of the best available figures when implementing that methodology. It must therefore be determined whether the BGN data used by the Commission to determine the final bid-ask spread constituted the best available figures in implementing the methodology defined by the Commission. In that context and having regard to the reasoning developed by the Commission in the contested decision (recitals 903 and 918) and, in particular, that in which it rejected the arguments put forward by the banks concerned in the course of the administrative procedure (recitals 919 to 929), those banks cannot confine themselves to arguing that the data used by the Commission suffer from one or more shortcomings, but, on the contrary, must show that, within the framework of the methodology which that institution has lawfully determined, there are in fact data better than those used by that institution and that those data are in fact available (see, by analogy, judgment of 24 September 2019, HSBC Holdings and Others v Commission , T‑105/17, EU:T:2019:675 , paragraph 324). In that regard, it should be noted that Crédit agricole essentially submits that, in the light of a study which it did not produce before the General Court and which it itself admits related to a ‘limited sample’ of transactions, the Commission should have given priority not to the BGN data, but to the actual spreads at which the transactions were executed by that bank and which should have been calculated on the basis of the ‘limited sample’ referred to above ‘as an indication’, in order to determine an ‘appropriate’ factor for reducing the bid-ask spread. Such imprecise and insufficiently substantiated evidence is not capable of demonstrating that a complete and consistent set of data other than that composed of the BGN data used by the Commission was available, or a fortiori that it was better than that set. On the contrary, the fact that Crédit agricole based its criticism solely on a ‘limited sample’ of data tends to show that it - and a fortiori the Commission - was not in a position to identify and document precisely the transactions which it itself carried out. In that regard, the Court cannot fail to note, as is apparent from recital 926 of the contested decision, that, during the administrative procedure, Crédit agricole also indicated that, ‘in the absence of actual transaction prices making it possible to calculate the quote spread, it is appropriate to use index price spreads (such as BGN) which are both specific to the bond and specific to a given period’, that ‘an appropriate source for [those] data is the Bloomberg data provider, which provides, for each bond and for each day, the end-of-day bid and ask prices’ and that ‘the difference between these bid and ask prices provides, for each bond and for each day, an end-of-day price spread that is specific to each bond’, which the bank does not dispute. Moreover, although the study referred to by Crédit agricole coincides with the arguments put forward by that bank in the course of the administrative procedure, it is clear from recital 925 of the contested decision that the Commission rightly rejected them, essentially on the ground that, under the methodology proposed by that bank, because of the volatility of intraday prices, the purchase price of a bond cannot be compared with the purchase price of a reverse transaction executed on the same day. Finally, if Crédit agricole intended to base its arguments on the economic study produced in Annex A.10 to its application, it was incumbent on it to refer to the precise points of that study that supported the present argument and not merely to point to its existence in the introduction to its application under the heading ‘Background to the application’ and to refer to it in its reply in the context of its criticism that the Commission had erred as regards the reference period taken into consideration for calculating the fine imposed on it (first part of its fourth plea in law). Like Crédit agricole, Credit Suisse argues that, for the purposes of determining its adjustment factor, the Commission should have used not the BGN data but the actual spreads at which the transactions were executed by that bank. In support of its claim, Credit Suisse raises various criticisms of the BGN data, which must be examined in turn. From the outset, the criticism that the Commission disregarded the actual spreads at which the transactions were carried out by that bank in favour of the BGN data on the ground that another party to the cartel could not produce the same data must be rejected. That criticism stems from an erroneous reading of recital 918 of the contested decision. That recital states that the use of BGN data is justified by the fact that the data contained in the Commission’s file do not provide a sample which would have been sufficient in itself to ensure that all of the relevant maturities and issuers declared by the parties in their replies to the Commission’s requests for information were proportionally represented. Credit Suisse’s criticism that the way in which the BGN data were compiled was unknown, making it unclear what significance should be attached to them, must also be rejected. It is admittedly true, as Credit Suisse points out and as is apparent from the documents drawn up by the Bloomberg platform, that the way in which the BGN data are compiled is partly unknown. However, that fact cannot deprive the Commission of the possibility of using those data in the present case. First, as is apparent from recitals 33, 220 and 725 of the contested decision, the data provided by that platform constitute data to which traders give a great deal of credence, in that, as Credit Suisse points out, platforms such as Bloomberg collect market information and then relay it to all market participants. Secondly, as is apparent from footnote 1060 to the contested decision, which refers to a document drawn up by the Bloomberg platform, the BGN data constitute a source which is based on executable and indicative prices from several operators and which indicates the prices available in order to reach a consensus and which, as the Commission points out, is drawn up by a third party to the proceedings. Accordingly, it cannot validly be argued that, on the basis of the partly unknown nature of the way in which they are compiled, the reference data in the form of the BGN data cannot be used by the Commission, particularly where Credit Suisse has in no way referred to market platforms which provide more accurate or more relevant information than the Bloomberg platform. Credit Suisse’s criticism, based on point 1.8 of the second Compass Lexecon report, that the BGN data ‘appear to reflect initial quotations and price estimates rather than differences in transaction prices’, must also be rejected. The report does not show with sufficient certainty that the real purpose of BGN data was the one mentioned by the authors of the report, particularly without a bibliographical reference and against the definition adopted by the Commission on the basis of a document drawn up by the Bloomberg platform. In addition, it must be noted that the statement taken by Credit Suisse from the report is accompanied by the following editorial caution: ‘as such, we understand that’, which considerably limits the probative value of that statement, which is not otherwise substantiated. Lastly, as regards the criticisms that the BGN data do not take account of the impact of trading volumes on SSA bond prices or that they overestimate the spreads obtained by Credit Suisse when it trades with other market makers, it has to be concluded that they do not appear to be unfounded. However, that finding cannot, in the present case, deprive the Commission of the possibility of using the BGN data and, as a result, invalidate the amount of the adjustment factor calculated by using them. First, the Commission cannot be criticised for having used data which did not accurately reflect the actual spreads at which the transactions were carried out by Credit Suisse when, as is clear from recital 918 of the contested decision, which is not disputed by Credit Suisse, that institution did not have accurate data which were sufficiently representative and it was therefore obliged to use a methodology based on alternative data which were necessarily less accurate in order to reconstitute a proxy. Point 16 of the Guidelines provides that ‘where the figures made available by an undertaking are incomplete or not reliable, the Commission may determine the value of its sales on the basis of the partial figures it has obtained and/or any other information which it regards as relevant and appropriate.’ Secondly, it is clear from paragraphs 831 to 850 and 854 above that a methodology based on the transaction data of the banks concerned cannot be more appropriate than that adopted by the Commission in the contested decision. Consequently, in the context of the methodology established by the Commission, the legality of which was confirmed in paragraph 852 above, the BGN data must be regarded as the best available figures, particularly in the light of Crédit agricole’s statements on the reliability of those data for the purposes of calculating the adjustment factor, referred to in recital 926 of the contested decision and noted in paragraph 878 above. Consequently, the third argument in the third part of Crédit agricole’s fourth plea in law and the first argument in the second part of Credit Suisse’s third plea in law must be rejected. – Credit Suisse’s criticism alleging that its proxy was overstated due to the inclusion of notional amounts of other banks in its regard In the context of the third argument in the second part of its third plea in law, developed in the alternative on the assumption that the General Court should hold that the Commission was entitled to use the BGN data, Credit Suisse criticises the Commission for having overestimated the notional amounts which were retained in its regard, by wrongly attributing to it amounts negotiated by other banks, and rejects its arguments on that point in recitals 928 and 929 of the contested decision. Including in Credit Suisse’s notional amounts transactions in which the bank buys liquidity (liquidity sourcing) (by paying the price to another market maker) would result in Credit Suisse being attributed transactions for which it could not receive income. A market maker would only receive a margin on the transaction with a client. In that respect, Credit Suisse points out that the price paid by a reseller to its supplier is not added to the calculation of the value of sales. In addition, the inclusion in the notional amounts of liquidity purchase transactions, which would account for 37.8% to 41% of Credit Suisse’s total transactions, would result in the same transaction being recorded twice when the transactions are carried out between two banks which took part in the infringement at issue. By the present argument, Credit Suisse complains, in essence, that the Commission included in the notional amounts (and not in the calculation of the adjustment factor), which were used in its case, transactions relating to liquidity purchase, which, in its view, constitute supply transactions which should not be taken into account in the calculation of the value of sales. While it is true that, in the context of determining a ‘value of sales’ within the meaning of point 13 of the Guidelines, only the amount of the sales of the undertaking concerned, and thus the total price which that undertaking charged its clients on the relevant goods or services market, must be taken into account for the purposes of determining the basic amount of the fine (see, to that effect, judgment of 1 February 2018, Kühne + Nagel International and Others v Commission , C‑261/16 P, not published, EU:C:2018:56 , paragraphs 66 and 67), that cannot, however, be the case here, given the methodology adopted by the Commission for the purposes of determining the adjustment factor. Under the methodology adopted by the Commission, referred to in paragraphs 748 to 759 above, the Commission determined an adjustment factor corresponding to half of the average bid-ask spread for each bank, to take account of the fact that each purchase or sale transaction carried out by a bank represented only half of a combined SSA bond buy and sell transaction (recitals 897 to 899 of the contested decision). It follows, as is clear in particular from point 11 of the letter on fines and from footnote 1054 to the contested decision, that, because of the specific features of the trading of SSA bonds, the Commission sought to determine the proxy not by applying to the seller the entire bid-ask spread for each SSA bond sale transaction - which would have prevented it from including transactions relating to SSA bond purchases in the notional amounts of the banks concerned - but by applying half of the bid-ask spread for the transaction in relation to the seller and half of the bid-ask spread for the transaction in relation to the buyer. Accordingly, in applying such a methodology, the validity of which was established in paragraph 852 above, the Commission had no choice but to include in the notional amounts of each bank concerned not only its sales transactions but also its purchases, including those involving purchases of liquid assets which, moreover, will for the most part, if not all, lead to a subsequent sale transaction. Moreover, from an accounting point of view, the fact that the Commission has retained half of the average price spread applied to the total amount of the transaction is equivalent to retaining a total average price spread applied to half of the amount of the transaction, which rules out any double accounting of the notional amount of that transaction, first, by the buyer and, secondly, by the seller. Finally, in so far as Credit Suisse argues that the liquidity purchase transactions carried out with the banks which took part in the infringement at issue should not have been taken into account in its notional amounts, it is important to recall that, in the context of the Guidelines, the concept of ‘proxy’, 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 (see, to that effect, judgment of 9 July 2015, InnoLux v Commission , C‑231/14 P, EU:C:2015:451 , paragraph 50 and the case-law cited). It follows that the determination of the proxy implies taking into account all the transactions carried out on the market affected by the infringement, as is the case for the determination of the value of sales (see, to that effect, judgment of 1 February 2018, Panalpina World Transport (Holding) and Others v Commission , C‑271/16 P, not published, EU:C:2018:59 , paragraph 30 and the case-law cited), for each of the undertakings that took part in the infringement at issue, as follows from point 10 of the Guidelines. In view of the methodology adopted by the Commission, referred to in paragraphs 748 to 759 above, pursuant to which it determined the proxy for each bank concerned by multiplying its notional amounts by its adjustment factor determined on the basis of half of its own average bid-ask spread, it is logical that, for a liquidity purchase transaction carried out by a bank that took part in the infringement at issue from another bank that also took part in that infringement, the Commission took into account the notional amounts corresponding to that transaction for each of those banks, as it did for any liquidity purchase transaction carried out by a bank that took part in the infringement at issue from another bank that did not take part in it. That methodology certainly means that the notional amount exchanged in a given liquidity purchase transaction is taken into account for both the seller of the SSA bond concerned and its buyer. However, that twofold taking into account follows from the very principles governing the setting of fines under the Guidelines and, in particular, from point 10 thereof, applied to the specific context of the present case, where the Commission determined the proxy with regard not only to sales transactions but also to purchase transactions. Moreover, to exclude from the calculation of the proxy part of the transactions which indisputably fall within the scope of the cartel complained of would have the consequence 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 (see, by analogy, judgment of 11 July 2013, Team Relocations and Others v Commission , C‑444/11 P, not published, EU:C:2013:464 , paragraph 77). In recital 929 of the contested decision, therefore, the Commission was right to take the view that, in the context of the methodology which it had legally defined and which was intended to determine a proxy representative of the entire trading activity of the banks concerned and, in particular, of that of Credit Suisse, it had to include in the notional amounts retained in respect of that bank the transactions relating to purchases of liquidity, including those carried out between the banks which took part in the infringement at issue. Consequently, the third argument in the second part of Credit Suisse’s third plea in law must be rejected as unfounded. (2) Crédit agricole’s criticism of the Commission’s decision not to use 2014 as the reference period for calculating the proxy Pursuant to point 13 of the Guidelines, ‘in determining the basic amount of the fine to be imposed, the Commission … will normally take the sales made by the undertaking during the last full business year of its participation in the infringement’. In recitals 894 and 935 of the contested decision, however, the Commission decided, for the purposes of calculating the proxy, to depart from point 13 of the Guidelines. In that regard, the Commission indicated that it was ‘more appropriate to base the annualised sales proxy on the value of sales actually made by the [banks concerned] during the months corresponding to their respective participation in the infringement’ and that, ‘thus, the notional amounts traded during the period of individual involvement in the cartel [were] annualised by a factor in function of the duration of the individual participation, dividing the total [notional amounts exchanged by a bank during its period of participation in the infringement at issue] by the number of full months of participation and subsequently multiplying this monthly average by 12’. The Commission justified that choice on three grounds, namely the varying size of the market for SSA bonds, the high volatility of price spread during the infringement period and the different periods in which the banks concerned were involved in the infringement at issue. Furthermore, in recital 895 of the contested decision, the Commission stated that the annualised notional amounts exchanged reflected the economic importance of the infringement at issue and the relative weight of each undertaking in that infringement, irrespective of their individual period of participation. In the context of the first part of its fourth plea in law, Crédit agricole claims that, by taking an annualised average of the notional amounts of the SSA bonds exchanged by that bank during its period of participation in the infringement at issue and not just the notional amounts exchanged by that bank during 2014 - the last full year of participation by that bank in the infringement at issue - the Commission infringed point 13 of the Guidelines, the principle of sound administration and its duty to state reasons. In support of that submission, Crédit agricole argues in its application that, in the contested decision, the Commission did not specify either the nature or the extent of the high volatility of the bid-ask spreads. Nor did it explain why the use of an annualised average of the notional amounts of SSA bonds exchanged was necessary in the present case, particularly as its participation began after the period of high volatility in 2011. In addition, Crédit agricole considers that the high volatility of the bid-ask spreads referred to by the Commission did not justify departing from point 13 of the Guidelines. That should not have been the case given that, in Crédit agricole’s case, the difference between the annualised notional amounts and the notional amounts traded in 2014 was only 10%. In its reply, Crédit agricole adds that the Commission’s justification based on the different periods of participation in the infringement at issue by the various banks concerned is absurd and contrary to its decision-making practice. As regards, first of all, the argument that the contested decision does not state adequate reasons for the Commission’s decision to depart from point 13 of the Guidelines, it must be noted that recitals 894 and 935 of the contested decision, the substance of which is recalled in paragraphs 918 to 921 above, set out clearly and unequivocally the reasons for that choice, thus enabling the banks concerned to know the reasons for that choice and the Court to exercise its review (see, to that effect, judgment of 23 November 2021, Council v Hamas , C‑833/19 P, EU:C:2021:950, paragraph 74 and the case-law cited). It is admittedly true, as Crédit agricole argues, that, in the contested decision, the Commission did not substantiate the nature and extent of the high volatility of the bid-ask spreads on which it relies. However, it would be excessive to require a specific statement of reasons for each quantitative element on which the Commission’s reasoning is based, where that reasoning is, as found in paragraphs 765 to 788 above, clear and unequivocal (see, by analogy, judgment of 1 July 2008, Chronopost and La Poste v UFEX and Others , C‑341/06 P and C‑342/06 P, EU:C:2008:375, paragraphs 108 and 111 and the case-law cited). That is all the more true in the present case because, first, Crédit agricole raised various arguments before the Court calling into question the high volatility of the bid-ask spreads on which the Commission had relied and which the Commission was able to justify in the context of the investigation of the present case (see, by analogy, judgment of 1 July 2008, Chronopost and La Poste v UFEX and Others , C‑341/06 P and C‑342/06 P, EU:C:2008:375, paragraph 110). Secondly, it cannot reasonably be considered that the absence of details concerning the nature and extent of the variability of SSA bond bid-ask spreads deprived Crédit agricole, a player on the SSA bond market, of the ability to assess for itself whether the objective finding made in recitals 894, 899 and 935 of the contested decision concerning the high volatility of the SSA bond market over the infringement period was true or, on the contrary, erroneous. Consequently, Crédit agricole cannot validly rely on a failure to state reasons in the contested decision as regards the reasons which justified the Commission’s decision to depart from point 13 of the Guidelines and to take into account, not the notional amounts exchanged by that bank during 2014, but an annualised average of the notional amounts exchanged by that bank during the entire period of its participation in the infringement at issue. As regards, next, the appropriateness of the choice made by the Commission, it should be recalled that point 13 of the Guidelines authorises the Commission to derogate from the rule that, for the purposes of calculating the amount of the fine, the sales made by the undertaking concerned during the last full year of its participation in the infringement at issue are to be taken into account, provided that that institution justifies that derogation to the requisite legal standard (see, to that effect, judgment of 17 December 2014, Pilkington Group and Others v Commission , T‑72/09, not published, EU:T:2014:1094 , paragraphs 212 and 213 and the case-law cited). In the present case, in recitals 894 and 935 of the contested decision, the Commission justified its decision to derogate from the rule laid down in point 13 of the Guidelines on three grounds, which are noted in paragraph 920 above. As can be seen from paragraph 923 above, in order to contest the justification for the derogation from point 13 of the Commission’s Guidelines, Crédit agricole limited itself in its application to contesting the second ground given by the Commission, namely the high volatility of SSA bond bid-ask spreads during the infringement period. It follows that the first and third grounds, based on the variation in the size of the SSA bond market over the infringement period and the different periods of participation in the infringement by the banks concerned, must be regarded as definitively established and relevant in order to justify, in the present case, the derogation from point 13 of the Guidelines. In that respect, it should be noted that Crédit agricole, in its reply, did dispute the third ground put forward by the Commission. However, in the absence of having been submitted in the context of the application, such an argument, which is of a new nature, which is not based on matters of fact or law which have come to light during the proceedings and which cannot be understood as constituting the amplification of its argument relating to the high volatility of the spreads in the quotation of SSA bonds is out of time and therefore inadmissible, pursuant to Article 84(1) of the Rules of Procedure (see, to that effect, judgment of 14 May 2019, Commune de Fessenheim and Others v Commission , T‑751/17, EU:T:2019:330 , paragraph 60). Having made those clarifications, it should be noted that the evidence provided by the Commission following a measure of organisation of procedure, which concerns a sample of SSA bonds (EIB and CADES) with a maturity of 0 to 3 years and with a maturity of 7 to 10 years, shows that, during the infringement period, those SSA bonds were indeed subject to significant price spreads, particularly between the end of 2011 and the end of 2013. That circumstance, taken together with the undisputed fact that the size of the SSA bond market had varied over the infringement period, highlights the fact that the annualisation of the notional amounts - leading to account being taken, for the purposes of calculating the basic amount of the fine, of the entire infringement period of each bank concerned - was necessary in order to reflect the economic significance of that infringement, as the Commission points out in recital 895 of the contested decision. Taking into account only the last year of each bank’s participation in the infringement at issue would have led the Commission to underestimate the economic importance of that infringement, by failing to take into account the years 2012 and 2013, during which the infringement had the greatest impact, due in particular to the high volatility of the bid-ask spreads. Such a conclusion cannot be called into question by Crédit agricole’s argument alleging, in essence, that its participation in the infringement - which began on 11 January 2013 and ended on 24 March 2015 - took place during a period of lower volatility of the bid-ask spreads, which should have led the Commission not to derogate from point 13 of the Guidelines concerning it. Not subjecting Crédit agricole to the annualisation of its notional amounts would have led the Commission not to use the same reference period for calculating the basic amount of its fine as that used for the other banks concerned, which would have prevented it from imposing fines reflecting the relative weight of each of the banks concerned in the infringement at issue, as is clear in essence from recital 895 of the contested decision. In addition, given that 2013 and even the beginning of 2014 were marked by volatility in spreads which was admittedly much less marked than that of 2012 and 2013, but greater than that of the period from the beginning of 2014, calculating the basic amount of Crédit agricole’s fine on the basis of the notional amounts exchanged by it during 2014 alone would have resulted in an infringement of the principle of equal treatment, to the detriment of the other banks which took part in the infringement at issue. In such a scenario, Crédit agricole would have been the only bank whose basic amount of the fine was calculated on the basis of a year with low volatility in SSA bond bid-ask spreads, even though it also participated in the infringement at issue in 2013, a year with significant volatility in SSA bond spreads. Finally, Crédit agricole does not provide any evidence to establish that the annualised notional amounts do not, for reasons specific to it, constitute an indication of its true size and economic power, or of the scale of the infringement which it committed (see, to that effect, judgment of 11 July 2014, Esso and Others v Commission , T‑540/08, EU:T:2014:630 , paragraph 96 and the case-law cited). In that respect, the fact, emphasised by Crédit agricole, that the annualised proxy used for it differs by only 10% from the notional amounts traded by that bank during 2014 cannot suffice. Accordingly, it was without committing an error that the Commission was able, in the present case, to derogate from point 13 of the Guidelines in calculating the basic amount of the fines imposed on the banks concerned and, in particular, on Crédit agricole. Finally, as regards the breach of the principle of sound administration alleged by Crédit agricole, it must be noted that it is not supported by arguments independent of those put forward to criticise the Commission’s choice not to use 2014 as the reference period for calculating the proxy, so that it must be rejected, for the same reasons as those set out in paragraphs 931 to 946 above. In the light of the foregoing, the first part of Crédit agricole’s fourth plea in law must be rejected. (3) The criticisms relating to the gravity multiplier (points 20 to 23 of the Guidelines) In recitals 940 and 941 of the contested decision, the Commission applied a gravity multiplier of 16% to the infringement. It justified that gravity multiplier essentially by the fact that the infringement at issue covered the whole of the EEA and took the form of price-fixing agreements, collusive exchanges of information, market sharing and client allocation, conduct which is among the most serious restrictions of competition. In the context of the fourth part of Crédit agricole’s fourth plea in law and in the third part of Credit Suisse’s third plea in law, the applicants challenge the gravity multiplier of 16% used by the Commission. First, as regards Crédit agricole, that bank submits that the Commission was not entitled to apply a coefficient for seriousness common to all the banks which took part in the infringement at issue and should have applied a ‘more appropriate’ gravity multiplier to it. According to Crédit agricole, an individualised and lower gravity multiplier was required in order to comply with the principle of equal treatment and to comply with the judgments of 7 June 1983, Musique Diffusion française and Others v Commission (100/80 to 103/80, EU:C:1983:158 , paragraph 129), and of 8 July 1999, Commission v Anic Partecipazioni (C‑49/92 P, EU:C:1999:356, paragraph 90). According to those judgments, the Commission was obliged, when determining the gravity multiplier, to take account of the duration of the infringement at issue, the conduct - possibly minor - of the undertakings concerned, their role in establishing that infringement, the profit made by them and their respective sizes. Crédit agricole was involved in the infringement for less time than the other banks and did not contribute to its establishment. In addition, its trader was less active than those of the other banks. Lastly, Crédit agricole was a ‘very small player’ in the SSA bond market, which it only entered in 2012/2013. In that regard, it should be stated at the outset that Crédit agricole cannot usefully rely on the judgments of 7 June 1983, Musique Diffusion française and Others v Commission (100/80 to 103/80, EU:C:1983:158 , paragraph 129), and of 8 July 1999, Commission v Anic Partecipazioni (C‑49/92 P, EU:C:1999:356, paragraph 90), in the absence of those judgments having been delivered in actions brought against decisions adopted at a time when the calculation of fines imposed on the basis of Article 101 TFEU was governed by guidelines, by which the Commission limited itself in the exercise of its discretion. As is expressly clear from the judgments of 7 June 1983, Musique Diffusion française and Others v Commission (100/80 to 103/80, EU:C:1983:158 ), and of 8 July 1999, Commission v Anic Partecipazioni (C‑49/92 P, EU:C:1999:356), the calculation of fines is made on the basis of Article 15(2) of Regulation No 17: First Regulation implementing Articles [101] and [102 TFEU] (OJ, English Special Edition 1959-1962, p. 87) - reproduced, in essence, in Article 23 of Regulation No 1/2003 - and not on the basis of the Guidelines, and in particular points 20 to 23 thereof (see, to that effect, judgment of 16 June 2011, Team Relocations and Others v Commission , T‑204/08 and T‑212/08, EU:T:2011:286, paragraph 87). It follows that the judgments of 7 June 1983, Musique Diffusion française and Others v Commission (100/80 to 103/80, EU:C:1983:158 ), and of 8 July 1999, Commission v Anic Partecipazioni (C‑49/92 P, EU:C:1999:356), cannot require the Commission, when determining the gravity multiplier (points 20 to 23 of the Guidelines), to take account of factors other than the intrinsic gravity of the infringement at issue, which, in any event and as provided in the case-law (see, to that effect, judgment of 16 June 2011, Team Relocations and Others v Commission , T‑204/08 and T‑212/08, EU:T:2011:286, paragraph 87), were duly taken into account by the Commission at other stages of the calculation of the fine imposed on that bank. In addition, it should be borne in mind that, although, in the context of the calculation of fines imposed on the basis of Article 101 TFEU, the Commission cannot disregard compliance with the principle of equal treatment (see, to that effect, judgment of 14 May 1998, BPB de Eendracht v Commission , T‑311/94, EU:T:1998:93, paragraph 309), it follows both from point 22 of the 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 (see, to that effect, under the Guidelines on the method of setting fines imposed pursuant to Article 15(2) of Regulation No 17 and Article 65(5) [CA] (OJ 1998 C 9, p. 3), judgment of 12 November 2009, Carbone-Lorraine v Commission , C‑554/08 P, not published, EU:C:2009:702 , paragraphs 27 and 29; under the Guidelines, judgments of 12 December 2012, Novácke chemické závody v Commission , T‑352/09, EU:T:2012:673 , paragraph 58, and of 14 May 2014, Reagens v Commission , T‑30/10, not published, EU:T:2014:253, paragraph 240). Thus, the assessment of individual circumstances is, in principle, carried out not in the context of the assessment of the gravity of the infringement, that is to say, when the basic amount of the fine is fixed, but in the context of the adjustment of the basic amount in the light of mitigating and aggravating circumstances (see, to that effect, judgment of 11 July 2013, Team Relocations and Others v Commission , C‑444/11 P, not published, EU:C:2013:464 , paragraph 103). It is in that sense that points 19 to 22 of the Guidelines envisage the determination of the gravity multiplier for the infringement at issue and not for each undertaking that took part in it. In the present case, a gravity multiplier of 16% for an infringement such as that found in the contested decision cannot be regarded as inappropriate or disproportionate. As the Commission rightly pointed out in recitals 939 to 941 of the contested decision, the infringement at issue covered the whole of the EEA and took the form of price-fixing agreements, collusive exchanges of information, market sharing and client allocation, conduct which is among the most serious restrictions of competition. Moreover, a gravity multiplier of 16% is very significantly lower than the upper limit of the range referred to in points 21 and 23 of the Guidelines for restrictions of competition taking that form. The Commission indicated in points 21 and 23 of the Guidelines that it would generally use a proportion of up to 30% of the value of sales, but that, for infringements such as horizontal price-fixing and market-sharing agreements - the classification used in the contested decision - the proportion used would generally be ‘at the higher end of the scale’. Moreover, contrary to what Crédit agricole maintains, the gravity multiplier applied to it does not reveal any breach of the principle of equal treatment. The specific characteristics of that bank and the relevant features of its participation in the infringement at issue were duly taken into account at other stages of the calculation of the fine, in accordance with the Commission’s obligation to take account of the individual conduct adopted by the undertaking in question (see, to that effect, judgment of 11 July 2013, Team Relocations and Others v Commission , C‑444/11 P, not published, EU:C:2013:464 , paragraphs 99, 100, 105 and 106). Thus, the shorter duration of Crédit agricole’s participation in the infringement at issue, as well as the fact that it did not take part in its establishment - a factor which is closely linked to the duration of its participation in that infringement - led the Commission to apply a duration coefficient of 2.20 years to it, whereas that coefficient was 3.28 years for BofA, 4.75 years for Credit Suisse and 4.18 years for Deutsche Bank, as is clear from recital 951 of the contested decision. Similarly, the smaller size of that bank was taken into account in relation to the proxy, which, because of the smaller volume of notional amounts exchanged by that bank, was much lower than that of the other banks, as is apparent from recital 905 of the contested decision. Finally, the fact that, during the infringement at issue, Crédit agricole’s trader was less active than the traders of the other banks was not, in the present case, a factor to be taken into account by the Commission. In addition to the fact that Crédit agricole does not allege, nor a fortiori demonstrate, that it fulfilled the conditions of the third indent of point 29 of the Guidelines, it is settled case-law that passive forms of participation in an infringement, such as the presence of an undertaking at meetings at which agreements with an anticompetitive object were concluded, without manifestly opposing them, constitute complicity which may give rise to liability under Article 101 TFEU, since tacit approval of an unlawful initiative, without publicly distancing itself from its content or reporting it to the administrative bodies, has the effect of encouraging the continuation of the infringement and jeopardising its discovery (see judgment of 21 January 2016, Eturas and Others , C‑74/14, EU:C:2016:42 , paragraph 28 and the case-law cited). Accordingly, the Commission was able, without committing an error, to apply a gravity multiplier of 16% to Crédit agricole in particular. That conclusion applies equally to Credit Suisse and is not called into question by that bank’s argument, in which it limits itself to claiming, in a single point of its application, that ‘the fine [imposed on it] … significantly overstates the gravity of the infringement [on the grounds that] the Commission has not demonstrated that the Price Discovery Communications are anticompetitive by object (second part of the first plea in law)’, that ‘there was no overall plan pursuing the same common objective before and after February 2013 (first part of the second plea in law)’, that ‘[it was] not aware of - and could not have reasonably foreseen - the allegedly anticompetitive bilateral communications engaged in by other banks after February 2013 (second part of the second plea in law)’, and that ‘the Commission has not proven to the requisite legal standard that the conduct occurring after February 2013 was continuous (third part of the second plea in law)’. Those four criticisms correspond exactly to those made by Credit Suisse and rejected by the Court in the second part of its first plea in law and in the first, second and third parts of its second plea in law, to which Credit Suisse expressly refers. It follows that the fourth part of Crédit agricole’s fourth plea in law and the third part of Credit Suisse’s third plea in law must be rejected as unfounded. (4) The criticisms relating to the duration multiplier (point 24 of the Guidelines) In recitals 950 and 951 of the contested decision, the Commission stated that it had calculated the duration multiplier provided for in point 24 of the Guidelines on the basis of the number of days of participation by each bank concerned in the infringement at issue, as determined in recital 842 of that decision. The Commission thus applied a duration multiplier of 2.20 years to Crédit agricole on the basis of the duration of its participation in the infringement at issue of 804 days, and a duration multiplier of 4.75 years to Credit Suisse on the basis of the duration of its participation in the infringement at issue of 1 738 days. In the fifth part of Crédit agricole’s fourth plea in law and in the fourth part of Credit Suisse’s third plea in law, the applicants challenge the duration multiplier applied to them. (i) The criticism of the duration multiplier of 2.20 years applied to Crédit agricole By the fifth part of its fourth plea in law, Crédit agricole complains that the Commission applied a duration multiplier of 2.20 years on the ground that the bank participated in the infringement at issue from 10 January 2013 to 24 March 2015. As it has shown in the context of its first and second pleas in law, its participation in the infringement at issue after the end of 2013 or, at the latest, after the beginning of 2014 has not been established. The Commission ignored or neglected several decisive factors relating to the duration of the infringement. First, the reduction in the amount of evidence available from the end of October 2013 demonstrates the cessation of the infringement at issue well before the date adopted by the Commission, which is confirmed by the absence of any discussions between the Crédit agricole trader and the BofA trader after the latter’s transfer from Deutsche Bank to BofA during the first half of 2014. Secondly, the conduct at issue - or, at the very least, the exchanges of information - did not have an anticompetitive purpose and were too far apart during 2013 and early 2014. In that regard, it should be noted at the outset that the duration multiplier used to determine the amount of a fine imposed under Article 101 TFEU bears no relation to the gravity and, therefore, the anticompetitive nature by object or effect of the infringement which justifies the imposition of that fine. Consequently, Crédit agricole’s argument that the conduct at issue did not have an anticompetitive purpose - which was rejected above - must be set aside. As to the remainder, it must be noted that Crédit agricole’s arguments, alleging that the infringement at issue was not continuous and that the Commission was mistaken as to the end date of that bank’s participation in the infringement, have already been rejected in the context of the reply to its second plea in law. Nevertheless, it was found in paragraph 145 above that the Commission could not validly rely on the discussion of 10 January 2013 and that, as a result, the start of Crédit agricole’s participation in the infringement at issue had to be postponed until 11 January 2013. In view of the fact that the Commission’s calculation of the duration multiplier was rounded down to the second decimal place (in particular for Credit Suisse and Deutsche Bank) and that the reduction of one day in the duration of Crédit agricole’s participation in the infringement at issue - from 804 to 803 days - implies a reduction of the duration multiplier from 2.20 to 2.19, an error must be found in the calculation of the duration multiplier applied to that bank. Accordingly, the fifth part of Crédit agricole’s fourth plea in law must be upheld. (ii) The criticisms relating to the duration multiplier of 4.75 years applied to Credit Suisse In the fourth part of its third plea in law, Credit Suisse complains that the Commission applied a duration multiplier of 4.75 years on the ground that that bank participated in the infringement at issue from 21 June 2010 to 24 March 2015. As Credit Suisse demonstrated in the fourth part of its second plea in law, its participation in the infringement after August 2014 has not been established. As the fourth part of Credit Suisse’s second plea in law has been rejected, the same must apply to the fourth part of its third plea in law. (b) The criticisms relating to the adjustment of the basic amount of the fine imposed on Crédit agricole As regards the adjustment of the basic amount of Crédit agricole’s fine made by the Commission under Section 2 of the Guidelines, that bank complains that the Commission, in part of the fourth part of its fourth plea in law and in part of the second part of its fourth plea in law, first, committed errors of assessment in examining the mitigating circumstances and, secondly, breached the principle of equal treatment in determining the specific deterrence coefficient applied to it (point 30 of the Guidelines). However, it must be noted that, in the fourth part of its fourth plea in law, Crédit agricole does not develop any argument in support of the allegation of errors of assessment in the examination of mitigating circumstances and confines itself to criticising the Commission’s assessment of the gravity multiplier for the infringement at issue, as set out in recital 941 of the contested decision. Consequently, in accordance with Article 76(d) of the Rules of Procedure, the argument alleging errors of assessment in the assessment of mitigating circumstances, developed in the fourth part of Crédit agricole’s fourth plea in law, must be rejected as inadmissible. Consequently, in the context of the challenge to the adjustments to the basic amount of the fine, only Crédit agricole’s criticisms relating to the adjustment made for the specific deterrence coefficient (point 30 of the Guidelines) and alleging breach of the principle of equal treatment will be examined. In support of its claim, Crédit agricole argues that, in recital 964 of the contested decision, the Commission could not, without breaching the principle of equal treatment, apply to it and to BofA a specific deterrence coefficient under point 30 of the Guidelines - of 1.2 and 1.3 respectively - because of the size of their turnover - approximately EUR 55 thousand million and EUR 73 thousand million respectively in 2020 - while not doing so in the case of Credit Suisse, which, for that same year, nevertheless had a significant turnover of approximately EUR 21 thousand million. In that regard, the Court of Justice has already held that, as is apparent from point 30 of the Guidelines, the illegality of which is not, moreover, alleged by Crédit agricole, the fact that the turnover of an undertaking is particularly high compared with that of other undertakings which took part in the cartel justifies, in order to ensure that the fine has a deterrent effect, an increase in the fine according to the economic power of the undertaking concerned (see, to that effect, judgment of 17 June 2010, Lafarge v Commission , C‑413/08 P, EU:C:2010:346 , paragraph 102, and order of 2 February 2012, Elf Aquitaine v Commission , C‑404/11 P, not published, EU:C:2012:56, paragraph 86 and the case-law cited). Moreover, while, in a given decision, the Commission cannot disregard compliance with the principle of equal treatment in determining the specific deterrent multiplication factors (see, to that effect, judgment of 13 December 2012, Versalis and Eni v Commission , T‑103/08, not published, EU:T:2012:686, paragraph 324), the undertakings concerned cannot nevertheless criticise the determination of those coefficients on the grounds that there is no exact proportionality between those coefficients and the turnover of the various undertakings concerned (see, to that effect, judgments of 18 July 2013, Dow Chemical and Others v Commission , C‑499/11 P, EU:C:2013:482 , paragraphs 88 and 91, and of 19 December 2013, Siemens and Others v Commission , C‑239/11 P, C‑489/11 P and C‑498/11 P, not published, EU:C:2013:866 , paragraphs 302 and 303). In the present case, it is expressly clear from recital 964 of the contested decision that the specific deterrent multiplication factors adopted by the Commission in relation to BofA and Crédit agricole were determined on the basis of the relative size of the turnover of those undertakings. Furthermore, Crédit agricole does not dispute that, in 2020, Credit Suisse’s turnover, which was not mentioned in recital 964, but which Crédit agricole communicated in its request, was very much lower than its own and BofA’s turnover. Credit Suisse’s turnover was 3.38 times less than that of BofA and 2.62 times less than that of Crédit agricole, which means that, of the banks which took part in the infringement at issue, Credit Suisse had considerably less economic power than those banks. It follows that, in order to ensure that the fine has a deterrent effect on BofA and Crédit agricole, the Commission was entitled to apply a multiplier for specific deterrence to them and, without breaching the principle of equal treatment, not to apply such a multiplier to Credit Suisse. The second part of Crédit agricole’s fourth plea in law must therefore be rejected as unfounded. In the light of the foregoing, Crédit agricole’s fourth plea in law and Credit Suisse’s third plea in law are rejected as partly inadmissible and partly unfounded. It follows from all of the foregoing that the first plea for annulment relied on by Crédit agricole in Case T‑386/21 and all of the pleas for annulment relied on by Credit Suisse in Case T‑406/21 are rejected. Consequently, in Case T‑406/21, the first and second heads of claim put forward by Credit Suisse and seeking, in essence, the annulment in whole or in part of Article 1(d) of the contested decision and the annulment in whole or, in the alternative, in part of the fine imposed on it in Article 2(d) of that decision must be rejected. In so far as Credit Suisse does not ask the General Court to exercise its unlimited jurisdiction under Article 261 TFEU and Article 31 of Regulation No 1/2003 (see paragraphs 46 and 67 to 69 above), the action brought by that bank in Case T‑406/21 must be dismissed in its entirety. By contrast, it follows from an examination of the second and third pleas relied on by Crédit agricole that its participation in the single and continuous infringement found in the contested decision is established only as from 11 January 2013 and that the Commission therefore erred when it considered, in Article 1(c) of that decision, that that undertaking participated in a single and continuous infringement as from 10 January 2013. Consequently, and as is apparent from the examination of the fourth plea in law, the Commission infringed Article 23(3) of Regulation No 1/2003 when it considered, for the purposes of calculating the amount of the fine imposed in Article 2(c) of the contested decision, that the infringement committed by Crédit agricole had lasted 2.20 years. It follows that the first and second heads of claim put forward by Crédit agricole must be upheld in part and that, first, Article 1(c) of the contested decision must be annulled in so far as, in that provision, the Commission finds that Crédit agricole participated in a single and continuous infringement from 10 January 2013 and, secondly, Article 2(c) of that decision must be annulled in so far as it sets the amount of the fine for which Crédit agricole is liable at EUR 3 993 000. In so far as the illegalities found in Case T‑386/21 have an impact on the determination of the amount of the fine and in so far as Crédit agricole seeks, in its second head of claim, a reduction in the amount of the fine imposed on it, the consequences to be drawn from those illegalities must be examined by the Court under its unlimited jurisdiction. F. Crédit agricole’s submissions in relation to unlimited jurisdiction In the context of its second head of claim put forward in the alternative, Crédit agricole asks the Court to exercise its unlimited jurisdiction to reduce the amount of the fine imposed on it. When exercising its unlimited jurisdiction under Article 261 TFEU and Article 31 of Regulation No 1/2003, the EU Courts are entitled, over and above the mere review of the legality of the penalty, to substitute their own assessment for that of the Commission, the author of the measure in which that amount was initially fixed, in determining the amount of that penalty (see judgment of 21 January 2016, Galp Energía España and Others v Commission , C‑603/13 P, EU:C:2016:38 , paragraph 75 and the case-law cited) and, consequently, to cancel, reduce or increase the amount of the fine imposed. In order to determine the amount of the fine imposed on an undertaking under its unlimited jurisdiction, it is for the EU Courts themselves to assess the circumstances of the case and the type of infringement at issue. That exercise presupposes, pursuant to Article 23(3) of Regulation No 1/2003, taking into account, for each undertaking penalised, the seriousness of the infringement at issue as well as its duration, in compliance with the principles, inter alia, of reasonableness, proportionality, individualisation of penalties and equal treatment, without the EU Courts being bound by the indicative rules laid down by the Commission in its guidelines, even though those guidelines may guide it (see, to that effect, judgment of 21 January 2016, Galp Energía España and Others v Commission , C‑603/13 P, EU:C:2016:38 , paragraphs 89 and 90). In that regard, it should be noted that the Court found, in paragraphs 145, 551, 984 and 1005 above, that the Commission had erroneously taken 10 January 2013 as the date on which Crédit agricole began to participate in the infringement at issue, which meant that that date had to be postponed to 11 January 2013. Nevertheless, subject to the implications of that finding for the duration of that bank’s participation in the infringement at issue and for the duration multiplier applied to it, it rejected as unfounded the criticisms raised by that bank and directed against the other factors relating to the determination of the amount of the fine imposed on it, in particular those relating to the gravity of the infringement at issue. In that context, the Court considers that, in the exercise of its unlimited jurisdiction, it can endorse the legally well-founded assessments made by the Commission in recitals 868 to 949 and 952 to 971 of the contested decision. However, it is for the Court to determine whether, in the light of the methodology used to calculate the proxy, the reduction in the duration of Crédit agricole’s participation in the infringement at issue from 574 to 573 working days necessitates a change in the amount of the fine imposed on that bank. That reduction involves not only a reduction in the annualised notional amounts used by the Commission, but also an assessment of the consequences of deleting the data relating to 10 January 2013 on the calculation of the final bid-ask spread for that bank, on which the adjustment factor used for it depends directly. It is also for the Court to consider whether the duration multiplier applied to Crédit agricole, as set out in paragraph 985 above, requires the amount of the fine imposed on that bank to be modified. However, having regard in particular to the rule of rounding down to the third decimal place adopted by the Commission for the calculation of the adjustment factor (see paragraph 759 above), but also to the duration - albeit slightly reduced - of Crédit agricole’s participation in the infringement at issue and to the gravity of that infringement, the Court considers that there is no need to amend the amount of the fine imposed on that bank for the single and continuous infringement referred to in Article 1 of that decision for the period from 11 January 2013 to 24 March 2015. Accordingly, the amount of the fine imposed on Crédit agricole should be set at EUR 3 993 000. IV. Costs Under Article 134(1) of the Rules of Procedure, the unsuccessful party is to be ordered to pay the costs if they have been applied for in the successful party’s pleadings. Since Crédit agricole has been almost entirely unsuccessful in its claims, it must be ordered to pay the costs in Case T‑386/21, in accordance with the form of order sought by the Commission. Since Credit Suisse has been unsuccessful in its claims in their entirety, it must be ordered to pay the costs in Case T‑406/21, in accordance with the form of order sought by the Commission. On those grounds, THE GENERAL COURT (Fifth Chamber, Extended Composition) hereby: 1. Joins Cases T ‑ 386/21 and T ‑ 406/21 for the purposes of the judgment; 2. In Case T ‑ 386/21: – annuls Article 1(c) of Commission Decision C(2021) 2871 final of 28 April 2021 relating to a proceeding under Article 101 TFEU and Article 53 of the EEA Agreement (Case AT.40346 - SSA Bonds) in so far as it finds that Crédit agricole SA and Crédit agricole Corporate and Investment Bank participated in the infringement from 10 January 2013 to 24 March 2015, and not from 11 January 2013 to 24 March 2015; – annuls Article 2(c) of Decision C(2021) 2871 final in so far as it fixes the amount of the fine for which Crédit agricole Corporate and Investment Bank and Crédit agricole are jointly and severally liable at EUR 3 993 000; – fixes the amount of the fine imposed on Crédit agricole Corporate and Investment Bank and Crédit agricole in Article 2(c) of Decision C(2021) 2871 final at EUR 3 993 000; – dismisses the action as to the remainder; – orders Crédit agricole and Crédit agricole Corporate and Investment Bank to pay the costs. 3. In Case T ‑ 406/21: – dismisses the action; – orders UBS Group AG and Credit Suisse Securities (Europe) Ltd to pay the costs. Svenningsen Tomljenović Mac Eochaidh Martín y Pérez de Nanclares Stancu Delivered in open court in Luxembourg on 6 November 2024. V. Di Bucci S. Papasavvas Registrar President Table of contents I. Background to the dispute A. The SSA bond sector 1. Price of an SSA bond on the secondary market (a) Price formation on the secondary market (b) Two-way prices offered by market makers 2. Trading positions and trading risk 3. Exchanges between banks via brokers B. The proceedings giving rise to the contested decision C. The contested decision 1. Establishment of the infringement at issue 2. The imposition of fines 3. The operative part II. Forms of order sought III. Law A. The requests that certain data be omitted from the public domain B. The purpose of the actions brought by Crédit agricole and Credit Suisse 1. Crédit agricole’s action 2. Credit Suisse’s action C. The purpose of the infringement found in the contested decision D. Admissibility of the criticisms directed against the interpretation of the discussions analysed in the contested decision 1. The pleas of inadmissibility relating to Crédit agricole’s criticisms 2. The pleas of inadmissibility relating to Credit Suisse’s criticisms E. Crédit agricole’s and Credit Suisse’s claims for annulment 1. Crédit agricole’s third plea in law, alleging breach of the principle of the presumption of innocence 2. Crédit agricole’s second plea in law and Credit Suisse’s second plea in law, alleging errors in the classification of the conduct at issue as a single and continuous infringement (a) Preliminary observations (b) The existence of agreements and/or concerted practices of an anticompetitive nature (1) The discussions found to be anticompetitive (2) The anticompetitive nature of the discussions, which Crédit agricole and Credit Suisse are entitled to contest (i) The discussions that Crédit agricole is entitled to contest – The discussion of 18 January 2013 – The discussion of 31 January 2013 – The discussion of 15 February 2013 – The discussion of 19 March 2013 – The discussion of 21 March 2013 – The discussion of 24 May 2013 – The discussion of 3 June 2013 – The discussion of 10 July 2013 – The discussion of 25 July 2013 – The discussion of 6 August 2014 – The discussion of 12 March 2015 (ii) The discussions that Credit Suisse is entitled to contest – The discussion of 28 September 2010 – The discussion of 8 February 2012 – The discussion of 10 January 2013 – The discussion of 12 March 2015 – The email of 24 March 2015 (3) Conclusion on the discussions analysed in the contested decision (c) The existence of a single and continuous infringement (1) The single nature of the infringement found (i) The existence of an overall plan pursuing a single anticompetitive objective before and during February 2013 (ii) The continuation of the overall plan pursuing a single anticompetitive objective after February 2013 (2) The continuous nature of the infringement found (i) The third part of Credit Suisse’s second plea in law, alleging lack of evidence of the continuing nature of the infringement at issue after February 2013 (ii) The second part of Crédit agricole’s second plea in law, in so far as it seeks to demonstrate that the infringement was not continuous (d) The participation of Crédit agricole and Credit Suisse in the single and continuous infringement found (1) The first part of the second plea in law put forward by Crédit agricole, formally alleging that it did not contribute to an overall plan (i) Crédit agricole’s intentional contribution to an overall plan (ii) Crédit agricole’s knowledge of all the other conduct implemented or envisaged by the other banks concerned (2) The second part of Credit Suisse’s second plea in law, alleging errors as to its knowledge of the bilateral discussions between traders of other banks or as to the fact that it could reasonably have foreseen them (e) The duration of Crédit agricole’s and Credit Suisse’s participation in the single and continuous infringement at issue (1) The second part of Crédit agricole’s second plea in law, in so far as it seeks to demonstrate that its participation in the infringement found was not continuous (2) The fourth part of Credit Suisse’s second plea in law, alleging that the Commission failed to prove that it participated in any unlawful conduct after August 2014 (f) Conclusion on Crédit agricole’s second plea in law and Credit Suisse’s second plea in law 3. Crédit agricole’s first plea in law and Credit Suisse’s first plea in law, alleging errors in the classification of the conduct at issue as a restriction by object (a) Preliminary observations on the subject and merits of the applicants’ criticisms (1) Crédit agricole’s criticisms (2) Credit Suisse’s criticisms (b) The classification of the conduct at issue constitutes a ‘restriction by object’ (1) Preliminary observations (2) The degree of harmfulness to competition required to classify conduct as a ‘restriction by object’ (3) The Commission’s assessment of the conduct at issue (i) Criticism of the Commission’s assessment of the economic context of the conduct at issue – Crédit agricole’s criticism, alleging that it failed to assess the economic context of the infringement at issue – Crédit agricole’s and Credit Suisse’s criticisms, alleging the erroneous assessment of the economic context of the infringement at issue (ii) Crédit agricole’s and Credit Suisse’s criticism, alleging that the conduct at issue is not sufficiently harmful to competition (iii) The criticism alleging that the conduct at issue is, in essence, ‘justified’ – Crédit agricole’s and Credit Suisse’s criticisms, alleging pro-competitive effects of the conduct at issue – Crédit agricole’s and Credit Suisse’s criticisms of the ‘necessary’, ‘legitimate’, ‘beneficial’ or ‘inherent and essential’ nature of the conduct at issue (4) Conclusion on the classification of the conduct at issue as a ‘restriction by object’ 4. Crédit agricole’s fourth plea in law and Credit Suisse’s third plea in law, alleging inadequate statement of reasons and errors in determining the amount of their respective fines (a) Criticism of the basic amount of the fines imposed on Crédit agricole and Credit Suisse (1) The criticisms relating to the amount of the proxy (i) Credit Suisse’s criticism alleging that no reasons were given for the calculation of the fine (ii) Crédit agricole’s and Credit Suisse’s criticisms, alleging breaches of the Guidelines – Crédit agricole’s and Credit Suisse’s criticisms, alleging the Commission’s refusal to use a methodology for calculating the spread based on data from their own transactions – Crédit agricole’s criticism, alleging use of fixed SSA bond weightings for the entire infringement period and for all of the banks concerned for the purposes of calculating the final bid-ask spread for each of those banks – Crédit agricole’s and Credit Suisse’s criticisms of the use of BGN data – Credit Suisse’s criticism alleging that its proxy was overstated due to the inclusion of notional amounts of other banks in its regard (2) Crédit agricole’s criticism of the Commission’s decision not to use 2014 as the reference period for calculating the proxy (3) The criticisms relating to the gravity multiplier (points 20 to 23 of the Guidelines) (4) The criticisms relating to the duration multiplier (point 24 of the Guidelines) (i) The criticism of the duration multiplier of 2.20 years applied to Crédit agricole (ii) The criticisms relating to the duration multiplier of 4.75 years applied to Credit Suisse (b) The criticisms relating to the adjustment of the basic amount of the fine imposed on Crédit agricole F. Crédit agricole’s submissions in relation to unlimited jurisdiction IV. Costs * Language of the case: English. © European Union The source of this judgment is the Europa web site. The information on this site is subject to a information found here: Important legal notice . This electronic version is not authentic and is subject to amendment. BAILII: Copyright Policy | Disclaimers | Privacy Policy | Feedback | Donate to BAILII