Flavio de Carvalho Pinto Viegas & Ors v Rosana Falcioni Cutrale & Anor

In these proceedings, the claimants (who number over 1500 in total and are, or are the heirs or representatives of the estates of, Brazilian located orange growers) allege that the defendants participated in a cartel which, in breach of Brazilian law, created a monopoly over the market in Brazil for the acquisition of oranges for the production of orange juice concentrate thereby resulting in (a) the purchase price of oranges paid to the claimants over the relevant period being driven down and/or (b) production and logistic costs borne by the claimants being increased. The claimants are all members of one or other of two trade associations located in Brazil.

As currently pleaded, the cartel is alleged to have come into existence by no later than 7 January 1999 and continued in operation until no earlier than 24 January 2006. The claim form in Claim Number CL-2019-000603 (“Viegas Claim”) was issued on 27 September 2019 and the claim form in Claim Number CL-2019-000727 (“Sanches Claim”) was issued on 22 November 2019. As currently pleaded, the claims relate to conduct that is alleged to have taken place between 21 and 14 years prior to the commencement of the claims. The claims are delictual in nature and governed by Brazilian law. The limitation periods relevant to the claims are those established by Brazilian law. The source of both the cause of action relied on by the claimants and the limitation period that applies to them is Brazilian federal statute law.

This is the trial of a preliminary issue concerned with whether the claims by the claimants in each of the claims are statute barred. The defendants contend that all the currently pleaded claims are statute barred as a matter of Brazilian law. The claimants maintain that on a proper understanding of the relevant law this is not so and decisions by the highest Brazilian civil court (the Superior Court of Justice (“STJ”)) that suggest otherwise are not binding, are wrongly decided and should not be followed.

It was also to be the hearing of an application by the claimants to amend their master Particulars of Claimso as to allege that the alleged cartel continued until August 2017. Since that application may be affected by the conclusions reached on the preliminary issue, the parties agreed that this application should be determined only following the hand-down of judgment on the preliminary issue, subject to me resolving an issue of Brazilian law concerning when the limitation period for delictual claims causing continuous damage starts to run and whether the claimants’ claims are to be characterised as in respect of continuous damage. I address that point in a separate section at the end of this judgment.

Given that the issues that I have to determine in relation to the preliminary issue exclusively concern Brazilian law, the trial was concerned exclusively with hearing evidence from Brazilian law experts. I heard expert evidence adduced on behalf of the defendants from Professor Didier, who is Professor of Civil Procedural Law at the Federal University of Bahia and on behalf of the claimants from Professor Frazão, who is an associate Professor at the Faculty of Law at the University of Brasília.

It is worthwhile noting at the outset the purpose of such evidence, which was summarised by the Court of Appeal in MCC Proceeds v Bishopsgate Investment (No 4) [1999] CLC 417 at [23] and [24] as being:

“… the function of the expert witness on foreign law can be summarised as follows:

(1)

to inform the court of the relevant contents of the foreign law; identifying statutes or other legislation and explaining where necessary the foreign court's approach to their construction;

(2)

to identify judgments or other authorities, explaining what status they have as sources of the foreign law; and

(3)

where there is no authority directly in point, to assist the English judge in making a finding as to what the foreign court's ruling would be if the issue was to arise for decision there.

24.

The first and second of these require the exercise of judgment in deciding what the issues are and what statutes or precedents are relevant to them, but it is only the third which gives much scope in practice for opinion evidence, which is the basic role of the expert witness. And it is important, in our judgment, to note the purpose for which the evidence is given. This is to predict the likely decision of a foreign court, not to press upon the English judge the witness's personal views as to what the foreign law might be. … the expert witness is entitled to give opinion evidence in the absence of direct authority, but we would underline the restrictions which it places upon him. His role is to ‘predict’ what the foreign court would decide, and only in this sense should he say ‘what answer should be given’.” [Emphasis supplied]

This approach has been followed consistently – see Dexia Crediop SpA v Comune di Prato [2015] EWHC 1746 (Comm); [2017] 1 CLC 969 at [34]; Perry v Lopag Trust Reg [2023] UKPC 16; [2023] 1 WLR 3494 at [11], where Lord Hodge giving the judgment of the Board of the Privy Council emphasised that “(i)t is not sufficient for a party to identify a judgment of a foreign court of first instance which may be on point and assert that the task of the appellate court is simply to analyse that judgment…” and Banco Intesa Sanpaolo v Comune di Venezia [2023] EWCA Civ 1482; [2024] Bus LR 228 per Sir Julian Flaux C at [165].

In consequence, “[(i)]f there is a clear decision of the highest foreign court on the issue of foreign law other evidence will carry little weight against it” – see Yukos Capital Sarl v. OJSC Rosneft Oil Co [2014] EWHC 2188 (Comm); [2014] 2 Lloyds Rep 435 per Simon J (as he then was) at [27(4)]. Whilst the English court is not obliged to apply a foreign decision if it is satisfied on the whole of the evidence that the decision does not accurately represent the law – see Yukos Capital Sarl (ibid.) at [29] - a court “… must be astute to give full weight to that judgment before concluding that … in future the [relevant foreign] court confronted with this issue would diverge from that high authority…” – see Deutsche Bank AG London v. Commune di Busto Arsizio [2021] EWHC 2706 (Comm) per Cockerill J at [108].

It follows that where the issue of foreign law has been resolved by the highest relevant court in the jurisdiction concerned (here the STJ ), that will be followed by an English court unless the party contending otherwise can show that the foreign court concerned has misstated or misapplied its own law and would resolve the issue differently if having to decide the same issue in the future, or as it was put in para. 11 of the claimants’ skeleton submissions “… would take the opportunity to correct itself.”

Whilst this will always be a high hurdle to overcome, in my judgement it is likely to be particularly high where (a) there have been multiple decisions to the same effect by the relevant highest foreign court (here, the STJ) as to how the issue that arises is to be resolved and (b)although there is no doctrine of precedent in Brazil (which is a civil law jurisdiction) equivalent to that which applies in common law jurisdictions, the STJ was created for the purpose of ensuring uniformity of interpretation of non-constitutional federal legislation, which is the source of Brazil’s limitation law. The claimants’ case is that they have overcome this hurdle because the arguments relied on by the claimants had never been considered or considered in sufficient detail by the STJ. It is that issue that was the primary focus of the expert evidence and cross examination of the experts at trial. This involves a careful analysis not merely of the decisions of the STJ but of the arguments that were put to it, the documentation available to it and the findings of the lower courts to the extent they establish the context in which the decisions of the STJ should be read.

Each of the claimants and defendants sought to attack the expert witness called by the other on the basis that I should as a consequence of such attacks reject the evidence of one or the other in its entirety and having done so then proceed to decide the issues that arise by uncritically accepting the evidence of the witness who (depending on which submissions I accepted) was left standing.

With respect, I consider that approach to be unhelpful. Neither expert gave consciously misleading evidence. Each to a degree gave answers in the nature of speeches in relation to points that they felt strongly about. I consider that Professor Frazão was on occasion too enthusiastic in her attacks on the decisions made by the STJ with which she disagreed, and I conclude that on occasion she strayed away from her role in informing me as to the content of the relevant part of Brazil’s limitation law and too far into her strongly but genuinely held view that the law should not be as it has been articulated repeatedly by the STJ. Although the claimants submit that there was not any proper basis for the suggestion that Professor Frazão’s prior role within CADE (defined below) had inappropriately influenced the evidence she gave, I consider that Professor Frazão’s evidence was heavily influenced by her genuinely held views as to how courts in Brazil should be approaching antitrust claims as opposed to how in fact they have been approached. By the same token, I considered Professor Didier’s evidence in some respects unhelpful – his evidence on the distinction between stand-alone and follow-on claims was not as helpful or accurate as it could or should have been and his approach in relation to the continuing loss issue which I determine at the end of this judgment was not as helpful or well-informed as it should have been simply because he had not read and therefore could not assist me in a fully considered manner in relation to the STJ decision on which Professor Frazão relied.

In those circumstances, I have considered the safer approach to consider the issues that arise on an issue by issue basis, largely by reference to the substance of each of the judgments of the STJ relied on by the parties and considering the degree to which they are undermined as accurate statements of Brazilian law by the points the claimants maintain the STJ has not considered or has ignored.

The primary facts relevant to the issues I have to determine are largely agreed and are set out in the List of Agreed Facts. It is necessary to start by setting out some framework facts that are critical to an understanding of the chronology that follows.

At the start of the period with which this claim is concerned, the Brazilian Competition Authority was the Secretaria de Direito Econômico (“SDE”). It is common ground that, as pleaded at paragraph 20 of the master Particulars of Claim, following reforms that came into force on 29 May 2012, the Administrative Council for Economic Defence (“CADE”) took over the functions that had previously been assigned to the SDE. It will be necessary to refer to a number of centrally important documents in what follows where the SDE and CADE are referred to often without distinction for this reason. Nothing turns on the distinction in these proceedings at any rate for present purposes.

One mechanism by which competition investigations can be compromised in Brazil is by parties under investigation by CADE in cartel cases entering into cease and desist Agreements with CADE. These agreements are generally referred to by Brazilian competition law practitioners as TCCs. That is how such documents are referred to in this judgment. A TCC is an agreement between CADE (or its predecessor) and those under investigation by it, by which CADE agrees to halt investigations as long as the parties to the TCC comply with what has been agreed. The parties agree (and in any event I find) that the interpretation of TCCs is a question of Brazilian law, as is the true meaning and effect of the decisions of CADE in respect of potential competition law violations that it investigates. The other method by which those under investigation can protect themselves is by entering into a leniency agreement with CADE (or its predecessor). The difference between a leniency agreement and a TCC is that a leniency agreement confers immunity from criminal prosecution in respect of the events with which it is concerned but a TCC does not.

It is next necessary to consider the Brazilian court structure. First instance courts are divided into branches according to the subject matter of the proceedings and are organised regionally. There is a route of appeal from the first instance to a second instance court. The ultimate court of appeal in Brazil (other than in relation to constitutional issues not relevant to these proceedings) is the STJ. The STJ is divided into a number of panels each concerned with different specialisms. It is the final court of appeal in Brazil concerned with interpreting non-constitutional federal legislation and was created for the purpose of ensuring uniformity of interpretation of federal legislation. It is common ground that this necessarily includes ensuring a uniformity of approach in relation to judge-made legal exceptions to statutory provisions of apparently general effect. This is relevant to these proceedings because when limitation starts to run in competition law cases subject to Brazilian law depends upon the scope and effect of judge-made qualifications to the applicable provisions of the Brazilian Civil Code.

Finally in relation to the framework facts, it is necessary to note that the company controlled by the defendants was and is called Sucocítrico Cutrale Ltda (“SCL”). It is referred to in a number of the centrally important documents quoted from below as well as the individual defendants to the claims I am concerned with.

The initial complaint concerning relations between the growers on the one hand and their processor counterparties on the other was made to the SDE in July 1994. The SDE opened an investigation and on 2 January 1995 published its Notice to that effect in the Diário Oficial (broadly the equivalent of the Official Gazette in this jurisdiction). On or before 31 October 1995, 11 processing companies, including SCL entered into TCCs with the SDE and a Notice was published in the Diário Oficial on that date that recorded the execution of the agreement, but no other details were published.

On 9 September 1999, the Consumer Protection Committee of the Federal House of Representatives referred to the SDE a complaint by the President of the Brazilian Association of Citrus Growers (“ABRACITRUS”) regarding alleged anticompetitive conduct between orange juice processors in Brazil. On 27 September 1999, the SDE initiated another investigation into alleged anticompetitive conduct against various parties including SCL and a Notice to that effect was published that day in the Diário Oficial. It identified the targets of the investigation and what was to be investigated was described as being “… the existence of conduct that violates the economic order, … which involve: a) entering into or applying an agreement with a competitor; b) sharing the market; c) creating difficulties for the operation of a competing company or supplier; d) discriminating against suppliers by means of differential pricing…”. By 12 September 2001, the SDE had found no conclusive evidence that “… could authorise the finding of an infringement of the economic order, despite the finding of the presence of structural conditions, pointed out by the economic literature, for the existence of a cartel and the verification of some points that raise concern with regard to competition in the market analysed” – see the Opinion of the Secretariat for Economic Monitoring of that date referred to in paragraph 25 of the 16 February 2006 SDE Technical Note to which I refer in more detail later in this judgment.

Following further investigations which consisted largely of seeking further information from the investigated companies, on 12 January 2006, Mr Paulo Ricardo Soares da Cunha Machado, a former commercial director of Coinbra-Frutesp S.A. (one of the companies then under investigation by the SDE) entered into a leniency agreement with the Brazilian Federal Government pursuant to Articles 35-B and 35-C of Law No. 8,884/94 (the “Machado Leniency Agreement”). Shortly thereafter “dawn raid” type searches were carried out at the request of the SDE at various locations including the offices of SCL. This exercise was called “Operation Fanta”. This exercise was widely publicised – see the report on 25 January 2006 in Folha de São Paulo, a daily national newspaper, in which it was reported that:

“The SDE (Secretariat of Economic Law) yesterday launched a search and seizure operation at five major orange juice exporters in the interior of São Paulo to investigate allegations of cartel formation in the sector.

The operation was coordinated by the SDE, an economic protection and defence agency attached to the Ministry of Justice, with the support of the Federal Police and the AGU (Attorney General's Office).

The companies targeted by the operation were as follows Montecitrus, located in Monte Azul; Coinbra, in Bebedouro; [SCL], in Araraquara; Citrovita, in Catanduva; and Abecitrus (Brazilian Association of Citrus Exporters), in Ribeirão Preto. In addition to the companies, the PF also investigated the home of a CEO of one of the companies investigated in Ribeirão Preto.

The companies account for 80% of the country's orange juice exports. When contacted by Folha, the companies did not comment…

The SDE has been investigating the cartel allegations since 1999, when the case was opened. The aim of yesterday's operation was to collect documents and electronic files to prove the allegations against the companies.

The complaint being investigated by the SDE is that the exporting companies are imposing prices on the orange growers, as the companies under investigation are essentially exporters.

The SDE already has strong evidence of a cartel, including several meetings and agreements with the aim of fixing the price of oranges.” [Emphasis supplied]

It was apparent from this report that firstly the SDE was investigating alleged cartel activity by orange processors and secondly that SCL was one of those being investigated. Since growers entered and enter into contracts for the forward sale of their crops directly with processors (indeed it is alleged that was one of the ways in which the alleged cartel carried into effect its objects) any grower would know whether it had sold or been forced to sell to one of those identified as being under investigation and whether it had been forced to accept an uneconomic price or a sale on uneconomic terms without being able to negotiate with any processor other than the one with which it had contracted. On the following day, Operation Fanta was reported on in Estadão, another daily national newspaper rather more graphically:

“"Federal Police Operation raids four orange juice factories"

An investigation is looking into an accusation of cartel formation; even an Uzi submachine gun was found at Cutrale.

With moments resembling a police film, the Secretariat for Economic Law (SDE), and the Federal Police conducted a search and seizure operation at the premises of four large orange juice processing companies and the Brazilian Association of Citrus Exporters (Abecitrus) two days ago. The result was the seizure of 27 packages of papers and documents, several computers, and even weapons. According to sources involved in the operation, an Uzi submachine gun was found in a drawer in the directors’ office at Cutrale, in Araraquara (SP). The companies Citrovita, Montecitrus, and Coinbra were also targeted in the raid.

Since 1999, the SDE has been investigating accusations of cartel formation by these processors. They are suspected of entering into an agreement for the division of the orange purchase market and forcing the reduction of input prices. The operation by the Police and the SDE aimed to seize any documents proving any such agreement between the companies.”

The claimants are largely all members of a trade organisation that has been at the forefront of advancing this claim, the full name of which is Associação Brasileira de Citrocultores and which operates and is known generally and, in these proceedings, as “Associtrus”. Mr Viegas is its President. The Estadão article was reproduced on the Associtrus website.

Just over a month later, on 24 February 2006, the SDE published an order in the Diário Oficial, which expanded the scope of the inquiry being undertaken. It recorded the identity of those subject to the proceedings. It identified José Luis Cutrale, José Luis Cutrale Júnior and SCL amongst others as defendants in the process. The Notice refers to a “Technical Note”. This document is of potential importance because of the detail it contains. There is however a dispute as to what if any part of it was available to the public and when. For present purposes it is agreed that the SDE produced a version of the 16 February 2006 SDE Technical Note labelled “Public Version” and that a version of the 16 February 2006 SDE Technical Note which did not bear the label “Public Version” on it was available on the SDE’s website on 14 March 2008. It is safer to proceed on the basis that this version was not readily accessible and so not material for the purposes of assessing what information was reasonably available to potential claimants for limitation purposes.

Operation Fanta and the order published on 24 February 2006 are important because the publicity attaching to Operation Fanta and the publication of the direction on 24 February have together been treated by the STJ on a consistent basis as being the publicly available information that was sufficient to start time running for limitation purposes in relation to claims by producers against processors for losses alleged to have been caused by the alleged cartel activity. The claimants in these proceedings maintain that this is wrong as a matter of Brazilian law; that the decisions of the STJ that reach that conclusion are all wrongly decided and that time started to run only much later when the defendants entered into TCCs with CADE and CADE decided to discontinue its investigations. As I explain in more detail below, this argument is one that the STJ has consistently rejected.

On 10 August 2006, Mr Machado testified to the Public Labour Ministry. His evidence included the following:

“1.

[I] worked at Frutropie from 1981 to 1988, becoming the company's director of raw material supply. Frutropic was subjected to economic pressure by the cartel of the orange juice industry, formed at that time by Cutrale, Citrosuco, Cargill, Frutesp and Branco Peres, until it became unviable and was taken over by the Dreyfuss Group in May 1988. In order to make the company unviable, the cartel attacked its suppliers by artificially increasing the price of raw materials and by making it difficult and impossible to rent warehouses and ships, thereby preventing or hindering the export of the finished product.

…

3.

- As soon as the takeover took place, the cartel immediately began to buy oranges from Frutropic's orange suppliers, enticing them with prices higher than the market price. Despite this, Frutropic was able to buy raw materials and products and was invited to join the cartel for the following harvest.

4.

- The cartel imposed a market share on Coinbra, with an increase in relation to its participation at the time and established a ban on suppliers with whom it could negotiate. At that time, all fruit was sold on the tree. Then, as now, the citrus grower could only supply one industry, which was determined by the cartel. Exceptionally, if the citrus grower has more than one orchard, he can supply more than one industry. The rule is one orchard per industry.

…

6 - Fruteps, although a member of the cartel, did not comply with the instructions concerning its participation in the market and, in retaliation, the cartel made it impossible for it to sell its stocks because of the price difference between the price applied by the cartel and the price agreed with Frutesp.

7 - In 1992 or 1993 the cartel met and decided to divide up the Frutesp producers among its [illegible] members in order to make Frutesp, which had a large stock and was unable to compete for raw materials, once and for all unviable. The cartel meeting was attended by Patrício de Camaré, then president of the Coinbra group, and Reinaldo Roberto Sesma, then president of the Coinbra juice division.

8 - They presented the deputy with a list of producers to be sought for the purchase of their crops and the instructions to pay above the market price and without any limit on supply. At the time, a carton cost around two dollars and Frutesp producers were approached with offers of up to fourteen dollars per carton.

9 - The cartel's intention was to retaliate against the suppliers without buying the asset, but Coinbra decided to buy the industry, even without the cartel's consent, and managed to renegotiate a new stake. Coinbra's new share was not the sum of its pre-acquisition share and Frutesp's share. It now has a larger share than before the acquisition, but less than the sum of the two. In return, the cartel gave Coinbra-Frutesp the contract for the transport of bulk juice, which Coinbra did not have.

10 - After the acquisition, the cartel members met to rebuild the cartel's foundations, which had been shaken during the process. This recomposition became known as the "Garden Agreement", which was nothing more than the allocation of rural producers for each industry and the prohibition of encroachment on the territory or "garden" of each. Suppliers could only sell to one industry and each industry could only buy from its suppliers.

…

12 - The "Garden Agreement" was in force for the 1993/1994 harvests. At the end of that year, after the cartel had been reconstituted and all the imbalances had been corrected, a new agreement was signed, covering the 1995 to 1999 harvests, with the participation of Cutrale, Citrosuco, Bascitrus, Cargill, Coinbra-Frutesp, Citrovita, CTM, Canbuhy/Montecimis and Frutax.

13 - This new contract consisted of fixing each company's share of the market for a period of five years. The negotiation of this agreement even required an inspection of the factories in order to determine the installed capacity and to avoid an increase of this capacity during the negotiations. During the inspection, the price range to be paid for the raw material was determined. The agreement also provided that violation of the market sharing rules would be punished by the offender compensating the aggravated party by supplying a finished product at a low price, either in Brazil or abroad. The terms of the contract were monitored on a weekly basis. The percentage of crushing was fixed not only for the harvest, but also monthly, and the industry could not exceed the fixed percentage…

14 -… which forced it to sell its entire production under threat of economic sanctions. Cutrale still buys juice from Sucorrico to prevent it from selling on the market and shares it with the other cartel members. Citrovita bought Sucorrico, but the contract remains in force.

15 - Coimbra bought the oranges destined for Frutax and divided them among the cartel members, and the company's assets were divided among all the cartel members after an auction, the results of which had been previously agreed at Barbosa's office in Musnich Aragão on 1 August 1999.

16 - At the beginning of 2000, before the deponent left Coinbra Frutesp, the cartel evaluated the previous harvest with a view to a new agreement. This evaluation was carried out with the participation of Plínio Rossetti, the auditor of the cartel contract. Coimbra Frutesp's legal department took part in the negotiation of the new agreement, as there were concerns about the CADE investigation that had begun the previous year.

17 - The execution of the cease-and-desist agreement at CADE in December/94 had no effect on the cartel or its operation.”

On 5 September 2006, a criminal complaint was issued by the São Paulo Prosecutor’s Office against, among others, the first defendant, alleging an agreement between orange juice processors to divide the market between themselves. The complaint document relies on similar evidence and analysis to that contained in the Machado Leniency Agreement, Technical Note and Opinion of the Federal Prosecutor’s Office. It identified as defendants both SCL and the late Mr Cutrale (whose estate is the first defendant in these proceedings). It recorded the substance of the allegations being made as:

“According to the records, from 1993 to the present time (continuous offence), on different dates, through telephone conversations, e-mails and meetings in different places in the State of São Paulo, continuously and repeatedly, the aforementioned companies, through their legal representatives, more recently by the accused, developing industrial and commercial activities directed to the field of the production and commercialisation of orange juices, previously coordinated and with an unequivocal unity of purpose, they formed agreements, covenants, arrangements and alliances, as suppliers, aiming at - the artificial fixing of prices and quantities sold and produced to control the national market, and more specifically the State of São Paulo and the metropolitan region of São Paulo, and to control the distribution network and suppliers of industrial orange juice, to the detriment of competition.

It is also established that the companies, through and as a result of agreements, understandings, arrangements and alliances as suppliers, with the aim of - artificially fixing prices and quantities sold and produced; to control the market by controlling the acquisition of raw materials and production, to the detriment of competition, the distribution network and suppliers, carried out by the respondents, representing the respective companies, established illegal rules to manipulate the supplier market (producers) of the raw material - orange. They established an exclusive network of suppliers for each of the companies involved in the criminal activities, with pre-agreed prices. Each supplier could and would only sell to the given company because the others refused to buy, and the maximum price was fixed in advance.

…

With the certainty and awareness of being, together, market leaders, the said companies, through their legal representatives - respondents, holders of decision-making power in relation to commercial activity, approximately 90% or more, of the said market, used the agreements among themselves, with the aim of managing, in a continuous and permanent way, the formation and control of an industrial orange juice cartel, with the aim of controlling and dominating the market through the control of the production, prices, sales and supply conditions of the products they produce.

According to the rules established by the cartel, the sale of the production of fruit - the raw material - was to be directed to each company participating in the group, according to a pre-established distribution. As a result of the manipulation of the purchase of fruit, the cartel was also able to control the quantity of production and consequently the respective prices, always increasing them and according to its interests, or reducing production, deliberately, in order to increase the price - reducing the supply in order to increase the demand, thus criminally violating the natural laws of the economy of free competition and market supply/demand.”

The document then set out an extensive history drawn largely from what Mr Machado had said in his statement, a description of the “Cartel Organisation”, remarking that the first iteration of the cartel was an “… agreement drawn up by the companies in September 1993, which continued throughout the 1990s.” and referring to the making of a new agreement in May 1995, the purpose of which was to “… formalise the practice of market sharing on a more solid basis and to establish new rules for the cartel, with quotas and purchase prices for raw materials and the market, internal and external stocks and the distribution of orange juice, including auditing with monitoring during the harvest, adjustments by means of transfers of raw materials between participants and complementary adjustment mechanisms.” It refers to a “… draft agreement attached to the case file … entitled "General Provisions of the Agreement", which runs to 13 pages. This document lays down the general rules of the agreement on the sharing of the orange market which the juice industry has complied with for the 1995/1996, 1996/1997, 1997/1998, 1998/1999 and 1999/2000 harvests.”

The Criminal Complaint described the main objectives of the agreement as being:

“… to determine juice production based on the group's strategy, to establish rules for the purchase of fruit and to control the harvest. It contained a number of rules that continue to manipulate the market for the supply of oranges to the juice industry. However, the most important restriction on the citrus market was the establishment of rules for the purchase of raw material from fruit growers. Through this agreement, the companies in the group were able to establish strict rules to be followed by the industries regarding the value and form of contracts for the purchase of oranges.” [Emphasis as in the original text]

There is then set out the terms of the relevant agreement in detail and went on to state that its effect was that the “… maximum price offered for the fruit was set jointly by the cartel members, in violation of the rules of the market - supply/demand and free competition - so that the citrus growers, tied together as they were, had no choice but to sell the fruit at the maximum price set by the cartel.”; that the alleged cartelists purchased fruit “… solely and strictly at the maximum price set for purchases by the juice industry group. They thus did not follow the market rules, as they only offered the maximum price agreed between them. They did not offer any other alternative to the growers, since the agreement provided that no other company would offer them a higher price for the purchase of their fruit.” The Complaint described the Brazilian Association of Citrus Exporters (“Abecitrus”) as “… the main implementing body of the illegal agreement entered into by the companies…” and the cartel as being “…fully coordinated through the direct action of ABECITRUS…” that acted throughout “… as a kind of executive body and arbiter of the cartel, monitoring compliance with the agreement, receiving information on the companies' periodic production and determining the compensatory measures to be taken by the cartel members.” All this led the prosecutor to conclude the Complaint by stating that “… the accused, through the use of their respective managerial positions in the companies, have continuously sought to maximise their profits, often by means of individual and concerted actions that artificially distort the proper functioning of these markets…”.

On 5 October 2006, Mr Marcio Luis Borella, a former employee of Agrocitrus, a subsidiary of Cargill Citrus, testified to the Federal Prosecutor’s Office, describing the operations of the alleged cartel from 1992 to 2000, and its continued operation in 2003 to 2004. The evidence is detailed and lengthy. It includes a statement to the effect that the various steps he took on behalf of his then employer were “… to tie the grower to a single company…”; and that in deciding the purchasing strategy in any year each member of the alleged cartel “… had its market share, with Cutrale having more or less 25%, Citrosuco more or less 22%, Cargill and Coinbra more or less 14%...” The evidence goes on to record that collusive conduct to drive smaller producers into insolvent liquidation and that in:

“… 1999 the companies decided that 15 to 30% of the fruit should not be harvested; that this was achieved to the extent that the companies did not buy the fruit from the producers, since the harvest was effectively the responsibility of the producers themselves; that in 1999 the companies again underpaid for the fruit and the multiannual contracts already mentioned were penalised; that by penalised it is meant that the industries refused to take all the fruit ready to be harvested; that, as a result, approximately 30% of the fruit in that crop was lost, causing great losses to the producers;…

… that in 2000 there was a breach of contract by the industries; that this was due to threats made by the industries to the producers that there would be a delay in harvesting the fruit; that as a result, producers who did not agree to renegotiate their contracts obviously had a partial loss of harvest…

that, if the purchase of fruit by one of the industries exceeded its market share, it was obliged to process the surplus in another industry or to compensate the others in juice; that this was a normal procedure at the end of each harvest; that the reverse was also true, if a company did not reach its market share, it would receive fruit or juice from another company to complete it;”

In July 2006, the SDE had launched a public consultation concerning the terms of a draft TCC that was ultimately rejected as contrary to Brazilian law. Two points arise from this. Firstly Associtrus (acting by Mr Viegas, then as now the President of Associtrus and the lead claimant in the lead action in these proceedings) commenced public law proceedings in Brazil for the purpose of preventing the 2006 TCC from being signed and secondly on 9 November 2006 a 36-page opinion of the Federal Prosecutor’s Office was submitted to CADE concerning the proposed 2006 TCC (the “Opinion”). The points that arise from each of these responses is relevant for the purpose of showing the information that was available to those minded to commence proceedings against amongst others the defendants in relation to losses said to have been caused by the alleged cartel activities that I have referred to above.

I turn first to the public law proceedings commenced by Mr Viegas. In my judgement that provides little or no assistance on the issue that matters because whilst it contains a number of generalised allegations, it is short on detail and whilst it refers to the Machado Leniency Agreement referred to earlier concerning Mr Machado, he is not identified by name and the documents seized during Operation Fanta are said to be sealed. It does however demonstrate that Mr Viegas and by extension Associtrus considered that there has been in existence since the early 1990s an alleged cartel involving seven dominant orange processing businesses (SCL, Coinbra-Frutesp S/A, Citrovita Agroindústria, Montecitrus S/A and Citrosuco, Citrovita Agroindústria, and Montecitrus S/A).

Turning now to the Opinion, it records at page 4 that it had been alleged by or on behalf of the orange producers that the November 1995 TCC had never been respected, that they alleged that “… growers are unable to sell oranges to a processor other than the one to which they delivered the product in the previous harvest and that there is clear evidence of collusion between SLCC processors…” and at page 5 that:

“The petitions and the producers' correspondence, in turn, report the following in the SDE's summary of the facts (pgs. 2861) "(i) that there had been a cartel between the Represented Companies for more than 15 years; (ii) that they would act together in the market for the purchase of oranges from small growers and, in this sense, if a grower sold his crop to a particular company, in the following harvest that grower would not be able to sell to the other industries in the sector because they would always offer him prices lower than those paid by his usual buyer; (iii) the represented companies would act jointly in the sale of SLCC, which is the reason why small companies are unable to remain in the sector and are soon closed down or bought up by large industries in the sector; and (iv) the possible verticalization of the sector, with the acquisition by large industries of farms to grow oranges themselves."”

In relation to Operation Fanta, it was recorded at page 7 that:

“It is important to note that the search and seizure measures carried out on 24 January 2006 by bailiffs accompanied by federal police officers were granted by federal judges in São Paulo, Ribeirão Preto, Araraquara and São José do Rio Preto, at the request of the Attorney General's Office, on the basis of information provided to the SDE on 12 January 2006 by a person who had signed a leniency agreement, regarding the existence and continuity of the cartel, even during the term of the TCC signed by the SLCC processors in 1995, when they undertook before CADE to abandon all practices of concerted action in the purchase of oranges from growers.

In fact, as stated in the extensive technical note PGS. By means of the Leniency Notice No. 2859/2926, issued by the Department of Economic Protection and Defence - DPDE and approved by the Secretary of Economic Law, the beneficiary of the Leniency Notice provided information that "has a unique wealth of details and constitutes undeniable indications of a violation of the economic order". There is clear information on the cartel, which has already been denounced several times, as well as on its participants and their form of organisation and operation. In particular, it is established that much of the information provided corresponds to the allegations made in the numerous complaints received by the SDE, in particular from 1999 onwards, which are being investigated in the present case' (page 2887).”

The Opinion also summarised the evidence provided by the “beneficiary of the Leniency Notice…” (i.e. Mr Machado) supported by relevant documentation as being:

“(i)

Between 1993 and 2000, the beneficiary was an employee of one of the companies involved and participated in "numerous meetings with other companies in the orange juice market with the aim of agreeing on the rigid sharing of the fruit supply market" (page 2878);

(ii)

A first agreement was signed by the companies represented with the aim of manipulating the prices of oranges sold by the growers to the industries and establishing control over the limit of the commercial supply of each company;

(iii)

The cartel's manipulation of the market for the supply of oranges took the form of strict rules to be observed by all the undertakings concerned. These rules or standards concerned, in particular, the form of the contract with the growers, the amount to be paid for the oranges and the non-purchase by each company of a percentage of the fruit in excess of its respective market share;

(iv)

The represented companies shared the market for the supply of oranges, tied growers to them and set up a compensation system so that if one industry purchased fruit from a grower linked to another industry, it would have to pay the appropriate compensation. In the words of the leniency applicant: "If one of the industries purchased fruit from a grower who was not 'linked' to it in the combination made between the companies, the purchasing industry was obliged to compensate the company that had priority over the seller of the fruit by supplying it with the same quantity of oranges that it had purchased from that citrus grower who was 'linked' to the other industry" (Pg. 2882);

(v)

The control between the companies acting in a cartel in the purchase of oranges was carried out through a clearing house managed by ABECITRUS, an entity that "acted as the great executive body of the illegal agreement entered into by the companies" (Pg. 2880);

(vi)

ABECITRUS appointed Mr Plínio Rosset to supervise the purchase of oranges, monitor the production of the cartel companies, attend meetings and receive reports and requests;

(vii)

In 1995, the undertakings concluded a new agreement to share the fruit market, applicable to the 1995/1996, 1996/1997, 1997/1998, 1998/1999 and 1999/2000 harvests, which maintained the cartel's operational mechanisms already established, such as the form and price of the purchase of oranges, the linking of producers, the periodicity of the submission of the undertakings' production reports, the meetings and the compensation system;

(viii)

Other compensation mechanisms were also established, such as the supply of juice to the company that did not achieve the expected number of fruit crushes, including the transfer of orange juice from one company to another abroad.”

The Opinion records the applications by the processors for injunctions concerning the use of the material seized as part of Operation Fanta for reasons described in the Opinion as “quite obvious” – that isthat “… there is a strong likelihood that there are important documents that can prove and confirm not only that the represented companies acted in cartels throughout the 1990s, even ignoring a cease-and-desist agreement signed with CADE in 1995, but also that the cartel is still active today, as already indicated by some of the evidence contained in the documents themselves and added to others submitted with this Opinion.” It was concluded that the processors had not complied with the 1995 TCC “…which was to put an end to the nefarious concerted practice that has caused enormous damage to orange growers.” and that to permit the proposed new TCCwould be “… tantamount to granting [the processors] the right to continue, freely and even more vigorously, to exercise full control over the orange market, to control production, to fix prices, in short, to act as if they were the true "owners" of this producer market.”

The Opinion then went on to set out a summary of the evidence of continuity of the alleged cartel after 2000 at page 16 and following and the emergence of new evidence in the form of statements from former employees appended to the opinion and public testimony given at the Public Ministry of Labour. It would unnecessarily extend the length of this judgment if I was to set out at length all that material. Reliance is placed on the statements of two individuals described as “… former employees who held key positions in one of the companies represented and who spontaneously appeared before the MPF to talk about what they knew about the cartel under investigation in this administrative procedure…” which contained evidence of regular meetings between the managements of the relevant processors, the exchange of information between them for the purpose of carrying the aims of the alleged cartel into effect, coordinated buying strategies for similar purposes in line with agreed market share that owed little or nothing to fair and free competition and the supply of oranges between the processors inter se to maintain the agreed artificial market share.

On 23 January 2007, Antonio Carlos Favero and 50 other orange producers filed a claim against Louis Dreyfus Company Sucos S.A. (one of the companies alleged by the claimants to have participated in a cartel with SPL) (“Favero Proceedings”). The claimants in the Favero Proceedings include three who are also claimants in the Sanches Claim (José Antonio Ardengue, the estate of Norival Candido Ferreira and José Antonio Ruiz Sanches). On the same day, Adelia Virginia Fioreze Costa and 45 other orange producers filed a claim against Cargill Agrícola S.A. and Citrosuco Fischer S.A. – Agroindústria (two of the companies alleged by the claimants to have participated in a cartel with SPL) (“Costa Proceedings”). Twenty-one claimants in the Costa Proceedings are also claimants in the Sanches Claim. In each case the loss claimed in an amount “… corresponding to the difference between what was paid and what should have been paid by Orange Box since 1995, plus default interest as well as the amount corresponding to the loss of profits and the resulting damages, all to be determined by an expert…” – see paragraph 60 of the Particulars of Claim in the Favero Proceedings and paragraph 63 of the Particulars of Claim in the Costa Proceedings. This formulation is relevant to the continuing damage issue I consider at the end of this claim because it reflects what I consider to be the true nature of private law claims in competition law damages claims of this sort.

In each case, the Complaint, the Opinion of the Federal Public Prosecutor, Mr Machado’s testimony and Mr Borella’s testimony was filed by the claimants as part of the claims. This demonstrates that this material was available to the numerous claimants in those proceedings. In the Favero Proceedings, reliance was placed on the Complaint, which is described as having been “… received by the claimants in their capacity as members of Associtrus - Brazilian Association of Citrus Growers…” and is quoted at some length in paragraph 20 of the Particulars of Claim as well as on “… the information obtained through a leniency agreement signed between CADE - Administrative Council for Economic Law, SDE - Secretariat for Economic Law, Federal and State Public Prosecutors and…” Mr Machado - and the “… opinion of the representative of the Federal Public Prosecutor's Office at CADE…” The Machado Leniency Agreement resulted in:

“… the former director of the defendant [began] to report on the existence of a cartel and to provide documents in support of the alleged facts. (Documents to be requested)

17- It should be noted that the Leniency Applicant confessed to having actively participated in the cartel from 1995, without interruption, as a former director of one of the industries, until the day of his dismissal, still providing information that the cartel remained active even during the period in which a Cease and Desist Agreement (TCC) is in force, which has never been respected 4. Read the attached statement. (Document 06).”

Reliance was also placed on the evidence of Mr Borella as clarifying “… the modus operandi of the juice industries in order to maintain the supply and price control of…” the orange crops. The terms in which this and the Costa claim was pleaded is material to the knowledge that was available to enable claims to be brought.

In the Favero Proceedings, reliance was placed on the concentration of juice production in the hands of a small number of “industries” (a word used in many of the documents to connote one of the corporate industrial processors so that for example SPL is such an “industry”) as creating an oligopoly that is “… well described in the prosecution's complaint, a copy of which has already been attached.” There is significant economic analysis of the sort familiar to competition lawyers which suggests price movements that cannot be explained by ordinary market behaviour. While this is referred to throughout the Particulars of Claim, paragraph 34 sums up the key arguments being presented:

“As can be seen from the attached graph (Document 10), between 1970 and 1990, the average price of a box of fruit on the tree was US$ 4.50 (value updated by the CPI - Consumer Price Index). Since 1995, with the increasing concentration of the market, this price has been reduced to US$ 2.50 per box on the tree, a reduction which is incompatible with the reality of the market and which clearly indicates that the prices paid to producers have been subject to an artificial reduction, while the consumer market has continued to expand, both in terms of demand and in terms of price. It is important to note that during the same period, in addition to the enormous increase in costs due to pests, etc., and the loss of profitability due to the lack of financial conditions of the producers, the industries transferred to the producers the burden of harvesting and transporting the fruit. Despite this, the average price of a box of oranges has fallen to the equivalent of US$ 2.50.”

That claims had been issued in Brazil in January 2007 were reported in the Associtrus newsletter for January–February 2007. In one such article, Associtrus reported that it was “… estimated that more than 200 orange producers have already filed lawsuits for the cancellation of contracts based on allegations of cartel formation and practices of oligopsony.” This strongly suggests the general availability of information sufficient to enable an orange grower to have unequivocal knowledge that it had been the possible victim on unlawful anti-competitive activity and the identity of those responsible. The claimants in these proceedings maintain however that if such was the case it is nevertheless legally irrelevant for the reasons summarised earlier.

Thereafter, things continued much as before until on 7 November 2016, CADE proposed TCCs to those being investigated and ABECITRUS, SCL and ten individuals, including the second defendant, agreed with CADE’s proposal and entered into TCCs with CADE. There is a dispute as to the effect of this document to which I turn in detail below. However, in summary, the claimants maintain that the document is to be construed as containing an admission by the relevant parties sufficient to enable the claimants in these proceedings to bring what they characterise as a follow-on claim against the defendants. The defendants dispute that as the correct analysis of the document, submit that the question of how the TCCs are to be interpreted is a question of Brazilian law and that the claimants’ argument is one that has been consistently rejected by the STJ applying those principles. The claimants accept that is what the STJ has concluded but maintains that conclusion is wrong both because it ignores critical parts of what appears in the TCC and fails to take account of what as a matter of law must be included in a TCC if it is to be valid.

In so far as is material for present purposes the TCC provided as follows:

“Clause Two - Admission of participation in the conduct.

2.2.

In accordance with the requirements in the applicable legislation, the execution of this Term of Commitment implies the admission, by the COMMITTED PARTIES, of the facts described in the “History of the Conduct”, which is an integral part of this term as its Annex I below.

…

Clause Three - Obligations of the Committed Parties

3.1.

Pecuniary Contribution - The COMMITTED PARTIES undertake to pay the Diffuse Rights Defence Fund a pecuniary contribution in the amount of ninety-two million, five hundred and three thousand, four hundred and four reais and twenty-nine cents (BRL 92,503,404.29), in a single payment, within 30 days after the publication of the approval of this Term of Commitment in the Federal Government Official Gazette.

…

3.5.

Future Conduct - The COMMITTED PARTIES undertake to:

3.5.1.

Refrain, irrevocably and irreversibly, from engaging in any of the conducts investigated in the case files of the Administrative Case referred to in Clause One;

…

Clause Four — Stay and shelving of the Administrative Case.

…

4.2

Once the deadline for full compliance with all obligations provided in this Term of Commitment expires and once such compliance is verified, the Cases referred to in Clause One will be shelved in relation to the COMMITTED PARTIES.

…

Clause Six – Confidentiality of the documents

6.1.

It is noted that the Cases referred to in Clause One are processed under secrecy with access to the case files limited to the Defendants that will only be able to view them, without making copies, upon signing a term of commitment not to use the information outside the scope of the respective administrative case, in accordance with the judgement of the Federal Regional Court …including after the shelving of the administrative and judicial cases.

6.1.1.

It is understood that the documents contained in the case files must be used only by the Administrative Council for Economic Defence - CADE and for the purposes of the aforementioned Administrative Case, … Therefore, their disclosure or sharing, in whole or in part, with other individuals or legal entities, in Brazil or abroad, is forbidden, and the breach of the confidentiality duty will subject the breaching parties to administrative, civil, and criminal liability.

…

Clause Nine – Term of Effectiveness

…

9.2.

Once CADE attests to the performance of the obligations, as provided in this Term, the Cases referred to in Clause One will be declared shelved by CADE in respect of the COMMITTED PARTIES…

Clause Eleven - Publication

11.1.

The public version of this Term of Commitment will be disclosed at the time of its examination by CADE sitting en banc and it will be made public after its homologation, keeping the confidentiality of the terms of the negotiations and the Annexes hereto.

Clause Twelve: Confidentiality

…

12.2

The public version of this Term of Commitment will be disclosed at the time of its examination by CADE sitting en banc and it will be made public after its homologation, keeping the confidentiality of the terms of Clause 2.2 and the Annexes hereto.”

Annex 1 referred to in clause 2.2 included the following provisions said to be material to these proceedings:

“CONDUCT HISTORY No. 56/2016

ANNEX I

(ACCESS RESTRICTED TO THE COMMITTED PARTY OF THIS AGREEMENT AND TO CADE)

On November 23, 2016, Mr. ONOFRE CARLOS DE ARRUDA SAMPAIO, Brazilian, lawyer, … legal representative of SUCOCÍTRICO CUTRALE LTDA. (hereinafter referred to as “INDIVIDUAL COMMITTED PARTY”); JOSÉ LUIS CUTRALE JR., (hereinafter referred to as “2nd INDIVIDUAL COMMITTED PARTY”), jointly referred to as COMMITTED PARTIES, come before CADE’s General Superintendence to present the history of the conduct investigated … involving the Brazilian market for purchasing oranges for the processing of frozen concentrated orange juice.

I. Summary description of the conduct

The COMMITTED PART[IES] acknowledge that information gathered by their commercial teams in Brazil from the citrus market, especially from independent orange producers and traders, may have been possibly shared with competitors in the context of sectoral discussions on this market; the same may have occurred with equivalent information obtained from the market by its competitors during the period under investigation.

…

III. Participants of the conduct

To the best of the COMMITTED PARTIES’ knowledge, the following Legal Entities may have participated in the acts mentioned in this Conduct History (Footnote: ¹)…

…

IV. Association

The conduct described in Section 1 may have taken place during industry meetings at the Brazilian Association of Citrus Exporters (Associação Brasileira de Exportadores de Cítricos - ABRECITRUS) or in any other specific contacts in Brazil.

VI. Detailed Description of the Conduct.

The documents below are part of the case record and refer to competing companies, they are limited to the Brazilian market, and national territory, and have no relation to foreign jurisdictions or territories…

VIII. Conclusion

The COMMITTED PARTIES came before me, the Chief Superintendent Officer of the Brazilian Antitrust Authority, to acknowledge the conduct described above regarding the Brazilian market for the acquisition of oranges for the processing of frozen concentrated orange juice. This history contains a description of the facts that were reported to me.

This history of conducts is part of the Instrument of Consent Decree entered into by the COMMITTED PARTIES. It contains confidential information and may only be accessed, on a restricted basis, by the COMMITTED PARTIES, by CADE and by the other Represented Parties, with the exclusive purpose of exercising the rights to full defence and due legal process.”

CADE ratified the TCC on 23 November 2016, on 25 November 2016, a redacted version was published and on 29 November 2016, notice of the approval was published in the Diário Oficial.

On 28 February 2018, CADE’s Administrative Tribunal issued a decision, which unanimously endorsed the reasoned vote given by the reporting Commissioner, Paulo Burnier da Silveira, concluding CADE’s investigation. On 6 March 2018 the official order pursuant to the decision was published in the Diário Oficial. It is common ground, recorded in Paragraph 42 of the master Particulars of Claim, paragraph 48 of the Defence, paragraph 12 of the Case Memorandum and paragraph 33 of the List of Agreed Facts, that together, the decision, reasoned vote, and order constitute CADE’s final decision in the administrative proceedings (“CADE’s Final Decision”). The claimants maintain the Final Decision constitutes a condemnatory decision which entitles the claimants to bring a follow-on claim. The defendants dispute this because they argue that on a proper analysis the Final Decision is no more than a decision to discontinue the CADE proceedings on the basis that the relevant parties under investigation have complied with the terms of the TCC they had entered into with CADE.

As mentioned at the outset, the Viegas Claim was issued on 27 September 2019 and the Sanches Claim was issued on 22 November 2019. If the claimants are correct in their submissions as to the effect of the TCC and Final Decision, then these proceedings were brought within time. If they are wrong about that then the question of whether the claims have been brought outside the applicable limitation period depends on when the claimants had, or reasonably had the means of acquiring, sufficient knowledge to start time running against them in relation to the claims in these proceedings, which as they are presently constituted relate to events occurring between 7 January 1999 and 24 January 2006.

In 2002, the current Brazilian Civil Code (“BCC”) took effect by replacing the preceding code which had come into force in 2016. It dramatically reduced the limitation period applicable to these proceedings from 20 years. It is common ground that the limitation period applicable to these proceedings is the three-year period set out in Article 206(3)(v) of the BCC, which is the general limitation period that applies to delictual claims including claims for loss allegedly caused by unlawful anti-competitive conduct. This follows from Article 189 of the BCC, which provides that “… the right to claim … is extinguished by the … limitation periods referred to in Articles 205 and 206.” By Article 206(3)(v) of the BCC:

“The limitation periods runs

…

(3)

For three years in

…

(v)

claims for civil reparation;”.

As is apparent from the quotation above, Article 206 does not specify the date from which time is to run. Articles 197 and 198 address various postponements relevant to family, succession and personal competency issues that do not arise. Article 199 prevents time running either when there is a suspensory agreement in operation or “… while the period is not yet due.” By Article 200, time does not run during the pendency of a criminal investigation and by Article 202, the running of time may be suspended by court order pursuant to an application issued within time or by insolvency or succession processes or by an acknowledgement of indebtedness by a debtor. None of these provisions are alleged to be relevant to this case. The result is that the limitation period would apparently start to run from the date of the delictual act relied by the claimant.

This gives rise to obvious difficulties where a claimant could not reasonably acquire the knowledge necessary to at least commence a claim until after the expiry of the three-year period. This led to a judge developed actio nata rule under which time starts to run at the date when the victim could reasonably have been aware of the facts constituting the violation of their rights and the identity of the party liable– see paragraph 41 of the first Opinion of Professor Didier. Professor Frazão agrees in principle but submits that it was correctly summarised academically by Rachel Saab in her work entitled “Limitation: function, Assumptions, and Initial Term”; Belo Horizonte: Fórum, 2019 as being that “… the use of shorter limitation periods is normally linked to the adoption of subjective perspectives on limitation, such that the commencement of the limitation period is linked to “the date on which the holder of the right became aware of the harm, its authorship, and effects, or could reasonably have become aware of such facts, based on objective standards of action””.

As I have said already, the claimants’ primary case on limitation is that time started to run in respect of their claims only after the publication of CADE’s Final Decision so that what knowledge was required and when is on their primary case essentially immaterial. However, if that is incorrect then they dispute that they could reasonably have had the requisite knowledge in 2006 as the defendants maintain and as the STJ has consistently held to be the position. In that context, there is a dispute as to the degree of knowledge required by a claimant before time starts to run. Professor Frazão maintains in Section III of her second opinion that time starts to run only from the point at which the victim acquires “… unequivocal knowledge regarding the violation of their right, the authorship, and the extent of the damage…”. Professor Didier considers this overstates the requirement of the rule. He refers in paragraph 44 of his first opinion to a decision of the STJ published in 2022 in which the principle was described as being:

“… the actio nata theory consists of verifying, in each specific case, whether, based on the normative postulate of reasonableness, the creditor was or should have been aware of the birth of the claim, which must be ascertained in accordance with objective good faith and the objective standards of action of the average man,

…

Thus, although not exhaustive and not cumulative, it can be concluded that the following are criteria that indicate the tendency to adopt the subjective bias of the actio nata theory: a) the submission of the claim to a short limitation period; b) the finding, in the specific case, that the creditor was or should have been aware of the birth of the claim, which must be determined on the basis of objective good faith and the standards of action of the average man; c) the fact that we are dealing with civil liability for an absolute unlawful act; and d) the express legal provision imposing the application of the subjective system”.

I return to the subject matter and depth of the knowledge that this exception requires later in this judgment.

As will be apparent from the summary set out above, the claimants’ case is that the alleged cartel carried its mission into effect by imposing uneconomic prices on orange growers. The mechanism by which harvests of oranges are or were at all material times sold to processors was by a contract between the grower and processor by which the grower sold the year’s (or a number of future years’) harvests to a processor at a fixed price expressed to be a price per box of oranges. The cartel is alleged to have forced uneconomic terms on growers by dividing the market as to require growers to sell their annual production to one of the processors in the alleged cartel by precluding other members of the alleged cartel from dealing with the grower concerned.

The defendants’ submission is therefore that time starts to run in a case such as this on the later of when either the claimant entered into the contract with the processor by which the objectives of the alleged cartel are carried into effect or the date when the victim would have had available sufficient information to “… have been aware of the birth of the claim, which must be ascertained in accordance with objective good faith and the objective standards of action of the average man…”. This of necessity requires knowledge on the part of the victim “… of the harm, its authorship, and effects…” or that the victim “… could reasonably have become aware of such facts, based on objective standards of action…”. Logically time would start to run from the date a supply contract was entered into where the grower had the requisite knowledge or could reasonably have acquired such knowledge.

The defendants maintain that the claimants in these proceedings had the requisite knowledge or could reasonably have acquired such knowledge by no later than 24 February 2006, after Operation Fanta and the publication in the Diário Oficial of notice of the commencement of administrative proceedings against identified persons including the defendants. They submit that this has been the consistent position of the STJ over a number of years and is consistent with the commencement on 23 January 2007 of the Favero Proceedings and the Costa Proceedings, with the availability to Associtrus of the material referred to above and the assertions made repeatedly concerning the commencement of claims and the availability of the material necessary to commence claims as set out Associtrus’ various publications and website. This leads the defendants to submit that their case is consistent both with the settled view of the STJ as to the law and the documentary evidence summarised above.

The claimants’ primary case is that the claims now advanced in these proceedings are properly to be characterised as follow-on claims which are founded upon the Final Decision of CADE at the conclusion of its investigation into suspected cartel activity, with the result that time did not start to run until 6 March 2018 because it was only then that there was published in the Diário Oficial the official order pursuant to the decision of CADE’s Administrative Tribunal to conclude CADE’s investigation. As I have explained already, it is common ground that the decision, the reasoned vote of the reporting Commissioner and the order pursuant to the decision together constitute CADE’s Final Decision.

The existence in Brazilian competition law of a distinction between a stand-alone and a follow-on claim is established by the jurisprudence of the STJ. This is apparent for example from the STJ judgment of Andrighi J in SCL v. Neto 2023/0388472-7, where having referred to the distinction, the judge continued:

“This distinction has a major impact on the initial term for the calculation of the limitation period. The limitation period in follow-on actions commences with the final adverse decision rendered by CADE, as it is from that moment onwards that the individual becomes unequivocally aware of the violation of the right and its extent. On the other hand, setting the initial term of the limitation period in standalone actions requires a case-by-case examination of the moment when the holder of the right became effectively aware of its violation, as there is no interference from the administrative body or, at least, there is no adverse decision recognising performance of the wrongful act.”

The claimants’ primary case is that the decisions of the STJ relied on by the defendants do not assist because either the STJ failed to appreciate the difference between stand-alone and follow-on claims in those cases or that on proper analysis at least some of the judgments on which the defendants rely concern stand-alone as opposed to follow-on claims and so are distinguishable and immaterial for present purposes. The defendants maintain the claimants’ primary case is mistaken because a follow-on claim requires either a condemnatory decision by CADE (which the Final Decision is not) or an admission of actionable misconduct in the TCC recognised by CADE’s Final Decision, which the TCC in this case does not contain.

The claimants’ alternative case is that even if that is wrong, it cannot reasonably be concluded from the material relied on by the defendants that the claimants had the necessary knowledge to enable them to commence proceedings against the defendants earlier than the publication of CADE’s Final Decision so that on that test too time did not start to run until 6 March 2018. I question that analysis: if the defendants are correct in their submission that the TCC contains no relevant admissions of alleged cartel activity and CADE’s Final Decision is not condemnatory so as to enable the claimants to bring a follow-on claim, it is difficult to see how either can be relevant to an assessment of when the claimants had sufficient knowledge to bring a stand-alone claim.

Aside from that, the claimants accept that in the Favero and Costa Proceedings, the claimants made use of material in the public domain but maintain that those cases are unhelpful to the issues that arise both because the material relied on in those claims was incomplete and inconclusive and also because some reliance was placed on material that was not “publicly available”. The claimants submit that there is no evidence that any claimants or members of Associtrus beyond the 95 individuals who joined the two claims played any role in, or had any detailed awareness of, the way in which the claims were ultimately formulated. The claimants also apparently place some reliance on neither case having yet been resolved. Relatively little if anything can be made of this last point – the progress that has been made appears to be the result of procedural and interlocutory delays. Neither claim has been struck out or dismissed either summarily or otherwise. The force of the submission that the material relied on in Favero and Costa was inconclusive depends on the depth of knowledge that is required for time to start running for limitation purposes in relation to stand-alone claims.

It is convenient to start by analysing each of the decisions of the STJ on which the defendants rely before then considering whether (as the claimants submit) I should treat them as wrongly decided. In doing so, it is necessary to remember that the defendants submit that (i) each of the decisions relied on concerned materially the same alleged infringements as in the present case; (ii) in each case the claimants alleged that they did not have the necessary knowledge to start time running applying the actio nata principle until publication of the CADE Final Decision in March 2018 and (iii) in each case the STJ rejected that submission, holding that the CADE Final Decision was not a condemnatory decision and the TCC did not contain any relevant admission so that they were incapable of supporting a follow-on claim and that for stand-alone claim purposes the relevant date for the commencement of the limitation period was either the date Operation Fanta took place and the publication in the Official Journal of notice of the administrative proceedings on 24 February 2006 (which was when a claimant could reasonably have become aware of the harm and the person or persons responsible for that harm) or (if later or after the time when the claimant concerned actually acquired that knowledge) the date on which the claimants entered into the relevant contracts with the processors and that in the result each claim was statute barred.

The defendants relied on eight judgments of the STJ at the trial. After the trial further judgments of the STJ were published that the defendants allege establish the same points that the eight originally relied on establish. I consider those later in this judgment, whilst recording at this stage that the claimants submit the post hearing judgments make no material difference because, they argue, they suffer from the same frailties that they allege the original eight judgments relied on by the defendants suffer from. It is also worth recording at this stage the point made by the defendants that by 12 June 2024, there had been no fewer than 39 decisions of the Brazilian courts that held cases on substantially similar facts to be statute barred applying the principles set out in the eight cases on which the defendants principally rely. It is a significant feature of this case therefore that the claimants are seeking from the Courts of England and Wales a conclusion concerning Brazilian limitation law as it to applies to antitrust claims that the highest relevant court in Brazil has never accepted despite multiple attempts by claimants to obtain such a ruling.

The eight relevant judgments on which the defendants place reliance are:

In this claim, issued on 22 November 2019, in respect of losses said to have been suffered as a result of the sale and purchase of oranges for the 2001/2002, 2002/2003, 2003/2004, 2004/2005, 2005/2006 harvests, the claimant (Mr Teles) sought a declaration that a forward sales contract entered into by him with SCL was void and damages for losses said to have been caused by “… the practice of abusive and illegal acts by the Defendant, together with other companies in the citrus sector, which would have carried out anti-competitive acts in the orange purchase market, through price control, which imposed on great losses on the Appellant, as well as his exclusion from the citrus sector.” The claimant relied on “… the administrative proceedings that were initiated at the Administrative Council for Economic Defence – CADE ended due to the execution of the Cease-and-Desist Agreement, in which scope the companies would have confessed to the unlawful practice from 1995 to 2006.”

The first instance judge dismissed the claim as statute barred. The claimants appealed, arguing that:

“… the counting of the time-barring statute of limitations could only have started when the “holder of the violated subjective right became aware of the fact and the extent of its consequences”. In this regard, he concluded that the claim could only have been brought after the Appellant had “possession of the evidence of the unlawful act, which are being processed in camera proceeding before the Administrative Council for Economic Defence (CADE)”.

As summarised in the judgment, the argument between the parties and the decision of the Court below was:

“While the Claimant claims that the initial term for counting the statute of limitations is the date of the final decision by the Administrative Council for Economic Defence (CADE) on 28/02/2018, the defendant argues that the Claimant was already aware of the alleged facts at the time of entering into the contracts.

…

The case law established an understanding that the limitation period starts from the decision that recognizes the practice of the unlawful act

…

In the situation of the case records, the decision ratified the Cease-and-Desist Agreement - TCC. However, there is no acknowledgment, by the Defendant, of the practice of the acts indicated by the Claimant in the opening brief, not even a decision by CADE recognizing the existence of a cartel, or confession by the Defendant in relation to this fact. Thus, in the absence of a decision by CADE on the practice of the fact narrated in the opening brief (formation of a cartel), the initial term of the limitation period is the one set forth in Article 206, §3, V, combined with Article 189, of the CC (awareness of authorship, enforceability of the right), which is the reason why the Claimant’s claim is time-barred.”

Before the STJ, the claimant submitted that “… through the signing of Cease and Desist Agreements (TCCs), the investigated companies confessed to the practice of the attributed conducts in exchange for the closing of administrative proceedings and that, even, it would have been foreseen, in the scope of the document, the payment of an administrative fine in case of subsequent conviction…” and “… the State of São Paulo Court of Justice contradicted the case law that determines that the initial term of the limitation period, in cases such as in the case records herein, occurs after the publication of CADE’s decision, regardless of whether its decision operative part is for the condemnation or for the closing of the administrative proceeding, provided that there has been recognition of the occurrence of the offense.”

The STJ dismissed the appeal. In giving judgment, Salomão J acknowledged that “… considering the moment of the injury as the starting point of the limitation period tends to be extremely unfair, ending up punishing the victim for negligence…” but nonetheless dismissed the appeal on the ground that:

“… CADE, the authority responsible for the administrative proceeding, did not characterize the conduct considered abusive as the formation of a cartel, and a Cease-and-Desist Agreement was signed as a condition for suspending the initiated proceeding, which was later extinguished, given the compliance with the obligations stipulated in the document.

9.

In the absence of a decision by CADE on the formation of a cartel, the limitation period is the one established in Article 206, § 3, V, of CC/2002 – three years –, and the initial term of its counting is the date of knowledge of the harmful fact – in this case, the moment of the execution of the contracts.”

The STJ concluded that the claim was to be treated as a stand-alone not a follow- on claim and time started to run from the date when the claimant was coerced into signing the sale agreements with the defendant because they were signed at a time when the claimant was or ought reasonably to have been aware of all the relevant facts sufficient to start timing running.

In the course of the judgment, the Judge explained the distinction between stand-alone and follow-on claims as being that in stand-alone claims, “…the victim presents the evidence of the alleged act, as well as the damage suffered…” whereas in a follow-on claim, “…the victim supports the entire claim on the evidence and decisions produced by the authority that judged and condemned the cartel”. Although Professor Frazão maintained that there was no valid distinction between a follow-on and a stand-alone claim other than that a follow-on claim gave rise to a different limitation period, I reject that evidence. There is no principled reason for such an approach. If such was the case then there would be no need for any test other than one based on time running from the moment the victim becomes aware of the damage, its extent and authorship of the injury with it being a question of fact in each case whether time started to run earlier than the date of the publication of a CADE final decision. The point concerning different limitation periods becomes relevant only if the stand-alone and follow-on claims are substantively different in the way described by the STJ Judge in Teles, who identified by reference to the writing of Mr Bruno Oliveira Maggi a substantive difference between the two types of claims in these terms:

“…claims seeking damages for competition damage may be based on anti-competitive conduct reported directly by the victims or conduct that was investigated by antitrust authorities. He explains that, in the first case, there is a standalone lawsuit, in which the victim presents the evidence of the alleged act, as well as the damage suffered. The second case is that of the follow-on claim, “in which the victim supports the entire claim on the evidence and decisions produced by the authority that judged and condemned the cartel” (Idem).

Maggi further states that the follow-on claims, in which the injured parties will only know that they are suffering damage when the cartel is revealed by the competent administrative authority, no chronological point that precedes CADE’s final decision could be considered as the date of the damage- event, nor would it be possible to determine its eventual consequences (damage-loss). In these cases, in which the action is based on a decision by the administrative agent, “it is clear that the milestone that determines the beginning of the limitation period is the verification of the damage-loss, which may or may not coincide with the damage- event and with the illicit act” (Idem).

The STJ in Teles concluded that the claim being considered could not be a follow-on claim based on the CADE Final Decision because “… the defendant company signed a Cease and Desist Agreement with the investigating administrative authority as a condition for suspending the administrative proceedings filed against it, which was later extinguished in view of the fulfilment of the obligations stipulated in that Agreement. Therefore, … the start of the limitation period, in these cases, cannot be the date of the condemnatory decision handed down by CADE, simply because there is no condemnatory decision.”

Professor Frazão accepted that when determining the appeal in Teles:

For reasons that I expand on below, I conclude that the STJ was entitled to reach the conclusion that it reached based on its analysis of the substantive difference between stand-alone and follow-on claims. The claim could not be a follow-on claim because the CADE Final Decision did not make findings against the defendants that it participated in any relevant alleged cartel and the TCC did not contain any such admissions. However, it remains to consider the two points relied on by the claimants – that construing the TCC in the manner it has been construed fails to take account either of the laws that regulate when CADE can enter into TCCs and the conclusions of the General Superintendent of CADE expressed when recommending that the proceedings against the defendants be discontinued, which it is contended formed part of CADE’s Final Decision. I return to that issue in detail below, but it suffers from the difficulty that there are no admissions in the TCC or findings by CADE that together could replace the finding that would otherwise have to be made concerning the existence of an alleged cartel if a damages claim is to succeed. Absent such findings it is difficult to see how a claimant could succeed in establishing causation and loss other than by proving the existence of the alleged cartel, which is what a stand-alone claim is concerned with.

Unless I am persuaded that Teles was wrongly decided applying the English law principles noted at the outset of this judgment, the STJ’s judgment appears to be “… a clear decision of the highest foreign court on the issue of foreign law…” that arises.

The claim was issued on 20 January 2020 against SCL and concerned losses said to have been caused by sale and purchase agreements entered into between 1999 and 2006. The first instance court dismissed the claim and the TJSP (the regional Court of Appeal) and STJ both dismissed appeals from that decision, again on the basis that time had begun to run from the date on which the claimants had executed the contracts by which effect was allegedly given to the objectives of the alleged cartel (between 1999 and 2006) and had become time-barred three years later.

The headnote summarises the point resolved by the STJ as being essentially the same as that resolved by that court in Teles – that is that “… the starting point of the limitation period for indemnity claims based on an alleged cartel, in the present case, is the execution of the contracts, insofar as the administrative procedure before CADE did not involve a formal acknowledgment of guilt…” The claimants’ case was that the limitation period “… should be counted "from the closing of administrative proceedings at CADE, on 06/03/2018 … or in the worst case from of the publication in the DOU on 27/03/2017 (pages 87) of the order of the superintendent who declared the Cease and Desist Agreement signed by the defendant with CADE to be fulfilled, at which time the appellants had unequivocal knowledge of the violated right and, consequently, the injury suffered and its full extent, as well as those responsible for the offense"”.

The Court of Appeal had rejected a submission to similar effect on the basis that the TCC did not contain or imply a confession by SCL. The STJ rejected the claimants’ appeal by reference to that submission applying the reasoning in Teles, which is quoted from at length, and it was held that limitation was triggered each time a forward sale contract was signed by the claimant. The last such contract was signed in 2006 and the court therefore concluded that “… the limitation period has long expired ...”. The courts at all levels had found that the claimants had the requisite knowledge of the harm, the authors of the harm and the extent of the harm as Professor Frazão acknowledged whilst maintaining that the conclusion that had been reached was wrong.

It is questionable whether time could be said to have started to run earlier than 2006 or the date when the relevant contract has been entered into whichever was later but that is not material given that the claim was started only in January 2020. The court rejected a submission that time started to run only following the CADE Final Decision on the basis of the reasoning in Teles and unless I am persuaded that Teles was wrongly decided applying the English law principles noted at the outset of this judgment, the STJ’s judgment in Jotto appears to be another “… clear decision of the highest foreign court on the issue of foreign law…” that arises.

This claim was issued on 3 March 2021 by an orange producer against a processor in respect of contracts for the sale of oranges between 2000 and 2010 in which it was held that time started to run from no later than the commencement of CADE’s administrative proceedings in 2006 and that in consequence the claim was statute barred.

In that case the claimant had alleged that “… the statute of limitations began to run with the ratification of the Cease and Desist Agreement, as it would have characterised full knowledge of the injury and its full extent.” The court below had rejected that submission because “… only a Cease and Desist Agreement (TCC) was executed, and no decision was made recognizing the existence of a cartel or admitting guilt by the defendant.” The lower court’s analysis as set out in the STJ’s judgment merits substantial reproduction in full:

“The statute of limitations commenced upon the claimant's awareness of the damage and the subsequent pursuit of reparations…

It is acknowledged that, at the time of contracting with the price adjustment, the claimant producer of oranges was undoubtedly aware of the market price and production costs. Furthermore, he did not receive a superior offer, which ultimately led to the conclusion of a deal with the defendant. However, there is no evidence in the case file to suggest that the claimant was aware of the existence of the cartel at the time of entering into the agreement.

It is clear, however, that the "Operation Fanta" launched in 2006, which investigated Brazilian companies in the orange juice industry and the formation of a cartel, as well as the initiation of the administrative proceeding before CADE published in the Official Gazette, received significant media coverage. In the same year (24/02/2006), the author became aware of the formalisation of discussions about the existence of a cartel in the citrus sector and the resulting losses.

From the moment the claimant was aware of the alleged existence of a cartel and the possibility of having suffered damages, their interest in claiming their right began. However, in this case, it was incumbent on the claimant to prove not only the causal link and the damage, but also the defendant's unlawful conduct.

…

In this case, it appears that the contract between the parties was in force for the period from 2000 to 2010. Despite all disclosure, the claimant maintained a contract with the defendant after becoming aware of the cartel. Therefore, the statute of limitations on the subsequent contract began with the date of the contract.

However, even so, the initial claim had long been covered by the statute of limitations. This is because, even if the date of termination of the contract in 2010 is considered, the present claim was only filed in 2021, which is well beyond the three-year term.”

The STJ concluded that re-opening these conclusions would necessitate a re-evaluation of the evidence which was not something that the STJ was permitted to do.

In my judgement, the conclusions of the STJ in Manzoni do not take the debate much further. However, the conclusion of the Court of Appeal from which Mr Manzoni was seeking to appeal to the STJ do merit examination because it sets out in clear terms the approach to be adopted in relation to stand-alone claims. Professor Frazão was cross examined about this case and although she said that when “… the STJ decides to not take cognisance on an appeal, it doesn't confirm or uphold the decision from a lower court, they say that they will not analyse. So there's no confirmation or ratification…” she also acknowledged that the Court of Appeal had found that time began to run from the date of each contract (a point that is not only relevant to the issue I am currently considering but also to the continuing damage issue that I turn to at the end of this judgment). The most that can be said of this decision is that there is nothing in it that suggests that the STJ considered Teles to be wrongly decided and the Court of Appeal had applied that decision without criticism.

This was a claim issued on 5 March 2021 (the day before the third anniversary of the publication of CADE’s Final Decision on 6 March 2018) by an orange producer against a processor alleged to be a member of the alleged cartel. Every court in Brazil up to and including the STJ concluded that the claims were barred three years after the claimant had signed the forward contracts with the defendant that were alleged to have been forced on the claimant as a result of the activities of the alleged cartel.

Before the STJ, the claimant had argued that “… it was only after the conclusion of the CADE proceedings that it became clear that the damage had occurred, since the question of cartel formation depended on the investigation carried out by the competent authority” and that:

“… The statute of limitations only began to run with the approval of the Conduct Cessation Agreement TCC, which, according to CADE's own internal regulations, "is signed only if there is an express recognition on the part of the company of its participation in the conduct under investigation, in accordance with Article 18[5] (Footnote: ²) of CADE's Internal Regulations...”

This is an important development because it is contended by the claimants in these proceedings that one of the reasons why the analysis in Teles and Jotto is flawed is because of the failure in either judgment to take account of Article 185 of CADE's Internal Regulations in construing the effect of the TCC and/or CADE’s Final Decision. Professor Frazão accepted that in this case Article 185 was drawn expressly to the attention of the STJ.

In dismissing the appeal as “without merit”, the STJ rejected the submission that account could be taken of CADE’s Internal Regulations because those regulations did not “… fall within the concept of "federal law" provided for in art. 105, III, "a", of the Constitution of the Republic.” The STJ recorded the finding of the local court that:

“Notwithstanding the situation faced by the claimant is not possible to consider the initial Term of the Cease and Desist Agreement TCC signed between the defendant and CADE (Administrative Council for Economic Defence) (pages 115/ 120). A careful reading of the TCC shows that there is no admission of guilt on the part of the defendant in the formation of a cartel and that it resulted in financial losses to orange growers.”

and that if that was to be revised that would require a re-examination of the evidence, which was not permitted in a special appeal, but in any event added that:

“Moreover, the Fourth Chamber of this Court has already ruled that,

"(...) if there is no CADE decision on the formation of a cartel, the limitation period is the one established in Art. 206, para 3, V, of the CC/2002 - three years - and the starting point of its counting is the date of knowledge of the harmful event - in this case, the date on which the contracts were signed" (REsp n. 1.971.316/ SP, Rapporteur Judge Luis Felipe Salomão, Fourth Panel, judged on 25/ 10/2022, Electronic Justice Gazette of 14/ 12/2022).”

It follows that in this case the STJ once again followed the reasoning adopted in Teles because the CADE Final Decision did not contain a decision on the formation of an alleged cartel and had ruled as legally inadmissible the attempt to rely on CADE’s Internal Regulations as relevant to an interpretation of the Final Decision and /or the TCC.

This was a claim also issued on 5 March 2021 by an orange producer against two processors, which was held by all courts up to and including the STJ to be statute barred because the proceedings were commenced in excess of three years after the forward sales agreements in issue had been entered into. Like Oliveira, it was a special appeal.

The arguments advanced on behalf of the claimant were in materially identical terms to those advanced in Oliveira. As in that case the claimant relied on Article 185 (Footnote: ³) of CADE’s Internal Regulations as requiring that “… a TCC is signed only if there is an express recognition on the part of the company of its participation in the conduct under investigation…”. It was contended there, and the claimants contend before me, that the effect of this provision is to require a Brazilian court to construe the CADE Final Decision and/or the TCC as an admission of participation in an alleged cartel. For reasons I explain in detail below that is not what Article 185 requires, and it does not lead to the conclusion that what appears in the TCC should be construed in that way. As in Oliveira, the appeal was rejected by the STJ as being “without merit” for substantially the same reasons as the appeal in Oliveira had been rejected.

Professor Frazão accepted that in this case Article 185 was drawn expressly to the attention of the STJ and the Court below but explained the failure by the court to accept the point being made as a failure by STJ to deliberate over or even consider the point. Later in her oral evidence she maintained that the conclusion was the result of a “… misinterpretation of the TCC…” She also considered the decision was wrong because a cartel was what she described as being a “… permanent infraction…” – the point I consider in detail below when determining the continuing damage point.

Whilst I accept that Professor Frazão would be entitled (if such was the case) to argue that the STJ had wrongly decided this case and those that had gone before because it had failed to take account of Article 185, I do not accept that I would be entitled to ignore a determination by Brazil’s highest relevant court on an issue such as this unless there were very good reasons for thinking that the court had misunderstood the argument and would adopt the opposite view if the issue were fully argued before it. That in essence is what is contended for by the claimants by reference to the failure by the STJ in later cases to admit an amicus brief submitted by CADE. I return to that issue below. This does not appear to have been something that was relied on in Ramos at any rate in either the STJ or the Court below it. Subject to these points, this decision would appear to be another example of the STJ following its own previous decision in Teles.

This was an appeal concerning another claim commenced on 5 March 2021 by an orange producer against two processors. In relation to limitation, it had been held that time started to run with the signature of the forward sale contracts because the claimant had all the relevant knowledge at that stage to satisfy the actio nata principle.

The Court of Appeal noted that various dates for the starting of time to run for limitation purposes had been postulated before concluding that “… the statute of limitations could not have started to run with the final decision of CADE (Administrative Council for Economic Defence), since it did not formally recognise the existence of the orange juice cartel, unlike other judgements analysing the cement cartel, in which there was a condemnatory decision.” This point is of importance to this case because it explains why there is a distinction drawn between stand-alone and follow-on claims. The latter is to be distinguished from the former because in the latter the findings of wrongful conduct do not have to be proved because the findings of the regulator (here CADE) are adopted and the claimant has only to prove causation and loss. The position concerning the cement cartel is wholly different from the Final Decision of CADE in respect of the alleged orange cartel because in respect of the cement cartel CADE found the existence of the cartel in its final condemnatory decision on 22/09/2015. The STJ has consistently held that was not the effect of the Final Decision concerning the alleged orange cartel and in reaching the conclusion it did in Fabbri, the Court of Appeal reached a similar conclusion as it explained:

“CADE's decision was only to close the case with respect to the defendant companies and other investigated parties due to the fulfilment of the respective TCCs, as well as with respect to the other investigated parties due to insufficient evidence, so that there was no recognition by CADE of the existence of the frozen concentrated orange juice cartel (SLCC), it means, no formal condemnatory decision... the TCC … does not imply an admission of guilt, unlike the Leniency Agreement, where guilt is a legal prerequisite … Therefore, the clear knowledge of the damage cannot be attributed to the approval of the TCC, an agreement which, strictly speaking, does not alter the pre- existing factual scenario … And so, I would like to stress that CADE's final decision would not start the statute of limitations.”

The Court of Appeal also added (in a judgment that merits being read in full) that the claimant had pleaded that "the price of the orange box has always been set unilaterally by the buyer, who, using the power to set the price thanks to the cartel, has always unfairly imposed values that are always below the price that the orange box would have achieved on the market if the cartel had not been formed..." and concluded that showed that:

“… the appellant was aware that it had been subjected to actions by the buyer that were not in line with market prices, a clear source of the right to claim compensation for alleged losses, not least because the proceedings at CADE did not directly involve the appellant and the nature of the right allowed it to act in its own interests from the outset.

Even if this were not the case, the decision that gave wide publicity to Operation Fanta and the existence of the proceedings at CADE date back to 2006 and constitute a notorious fact, with the actual possibility of knowledge of the imputation of practices by the defendant in the outbreak of the operation, and it is unreasonable to consider the long process of the proceedings at CADE as relevant.” [Emphasis supplied]

The appeal from the decision of the Court of Appeal was dismissed. Andrighi J gave the lead judgment. Having recorded the appellant’s argument that the starting date for the running of time should be the date of execution of the TCCs, she then analysed the law and its application to the orange cartel allegations in terms which it is necessary that I reproduce in full:

“3.

Regarding the controversial issue in the case file, it is recalled that in the Brazilian Antitrust System, especially with regard to the claim for individual redress for competitive damage, two types of action coexist: (i) follow on and (ii) stand alone, which are distinguished exactly by the actions of the specialized administrative body in the area, the Administrative Council for Economic Defence - CADE. That is, when the alleged violation of economic standards depends on a decision by CADE, it will be a follow-on case, and, on the other hand, actions whose illegality was not decided by the specialized body will correspond to stand alone, being found and reported in court directly by the victims.

4.

This distinction has a major impact on the initial term for calculation of the limitation period.

5.

This is because the limitation period in follow on actions commences with the final adverse decision rendered by CADE, or the homologation of the Cessation of Conduct Agreement which recognizes the wrongfulness of the ceased practice, as it is from that moment onwards that the individual becomes unequivocally aware of the violation of the right and its extent. In this sense, when the requested redress has as its cause of action the administrative acknowledgement of an existing cartel whereby the victim bases "its claim on evidence and decisions issued by the authority in charge of investigating the existence of a cartel, which in the Brazilian case is the CADE", we are facing an action of the follow on type…

6.

On the other hand, setting the initial term of the limitation period in stand-alone actions requires a case-by-case examination of the moment when the holder of the right became effectively aware of its violation, as there is no interference from the administrative body or, at least, there is no adverse decision recognizing performance of the illicit act. Likewise, this type of claim applies when a Cessation of Conduct Agreement in which signatories only undertake to interrupt possible suspicious practices, without there being any recognition of a illicit act the administrative body.

7.

In this context, recent decisions of this Superior Court, when examining special appeals also arising from the infamous Citrus Cartelization ("Operation Fanta"), concluded that the initial term of the limitation period corresponds to the moment in which the victims, on a case-by-case and individual basis, became unequivocally aware of the allegedly criminal practices, not automatically binding them to the TCC homologation date.

8.

For example, the Fourth Panel decided in Special Appeal No. 1.971.316/SP, DJe 12/14/2022, that the "conduct deemed abusive were not characterized as formation of cartel by CADE, the authority responsible for the administrative case, and a Cessation of Conduct Agreement was executed as a condition to suspend the commenced proceedings, which was subsequently dismissed, as the obligations set in the document were complied with". In the judgement of the appeal, the execution date of the orange purchase and sale agreement was set as the initial term for the limitation period, as it corresponded to the moment when claimant became aware of the allegedly unlawful conduct.

9.

In another action in this Court, it was decided that "the moment of unequivocal knowledge of the harmful fact took place with publication of the commencement of the administrative case at CADE, for investigation of a cartel formation on 02/24/2006, a fact widely publicized by the media, with the launching of the Operation Fanta" (AgInt in AREsp No. 2.322.612/SP, Third Panel, DJe 09/13/2023).

10.

This same line of reasoning was adopted in the vote cast in the Special Appeal 2.133.992/SP, of which I am the rapporteur, with the judgement initiated before this Third Panel on 10/15/2024 - and the case was held under advisement with a request for examination by the Hon. Justice Cueva. In that case file, in line with the previous judgments of this Court, I held that the claim for redress of competitive damage, where resulting from homologation of a Cessation of Conduct Agreement without recognition of the practice of the illicit act before CADE, constitutes a stand-alone action and, as such, does not have its initial term of the limitation period linked to the confirmatory decision of the administrative body. In fact, as there is no recognition of the wrongful practices by the signatories, the initial term of the limitation period must be the date on which the victim becomes aware of the conduct that harms them and may take legal action to remedy and/or redress it.

11.

On the other hand, through the opinion presented in this special appeal (Special Appeal 2.166.984/SP), the Hon. Justice Cueva argues that the initial term of the limitation period, even in case of a stand-alone action, is the date of publication of the CADE decision that recognized compliance with the requirements set in the Cessation of Conduct Agreement. In summary, it is alleged as follows: "although the signature of the TCC does not imply a confession of fault, such characteristics do not annul the fact that the signature of the instrument implies the recognition of the committed party's involvement in the conduct under investigation (Art. 185 of CADE's Internal Rules). In other words, involvement in the conduct is recognized; hence, it is reasonable to understand that the holder of the violated right will only have unequivocal knowledge of the harm from this milestone onwards, which should start the limitation period" (page 8).

12.

However, with all due respect, I disagree with the Rapporteur of this special appeal and affirm the content of the opinion issued in Special Appeal 2.133.992/SP. It is therefore reiterated that stand-alone actions do not depend on a decision in the administrative case initiated before CADE and, precisely because of this independence, the individual may already file the claim for redress before the administrative body's statement, directly proving in court any action that is not in compliance with the Brazilian Antitrust System.

13.

It is therefore confirmed that the Cessation of Conduct Agreement signed with CADE does not imply a confession of the practice of a wrongful act. At most, it entails the recognition of the committed party's involvement in the conduct, which - I repeat - does not entail confession of fault nor recognition of the wrongful conduct.” [Emphasis supplied]

The judge then returned to the reasoning of the lower court reproduced above, which she quoted at length with implicit approval and then concluded that on the basis of “these reasons” (which I conclude to be both hers as set out in the judgment and those in the judgment of the court below which she reproduced in her judgment) she dismissed the appeal.

Four out of the five judges sitting agreed with Andrighi J. The dissenting judge (Justice Cueva) drew attention to a clear distinction between the contents of a TCC and a leniency agreement that I summarised earlier. He concluded that “… even though the signing of the TCC does not imply a confession of fault, such characteristics do not nullify or weaken the fact that the instrument implies the recognition of participation in the conduct investigated by the party making the agreement (Art. 185 of CADE's Internal Rules). In other words, involvement in the conduct is recognized (which is quite different from recognition of fault) …Therefore, the fact that the economic agent acknowledges in the TCC their participation in the conduct does not automatically imply a confession of fault, the existence of an unlawful act or the duty to compensate.”

Professor Frazão was cross examined about Cueva J’s observations set out above - see T3/79, in the course of which she accepted that “…Justice Cueva is crystal clear that the signing of a TCC does not imply a confession of infraction. That's what's required for it to ground a follow-on action…” – see T3/794-7 – but asserted that this judge (described by Professor Frazão as “… the best competition specialist judge in the highest civil court in Brazil…” – see T3/79/6-10) had proceeded from a mistaken premise (that the claim was a stand-alone claim) – see T3/80/16-19. I consider this misplaced for the reasons given above. The reasoning of the majority is clear – the claim is not a follow-on claim because the Final Decision of CADE does not contain findings or admissions and therefore any claim must by definition be a stand-alone claim (because liability cannot be established by reference to CADE’s findings alone) with the principles concerning when time starts to run depending on when the claimant in any particular case had all the relevant knowledge to satisfy the actio nata principle. It may well be that a claimant is entitled to point to the sort of concession made in the TCC in this case and rely on it as part of the evidence it uses to prove the existence of an alleged cartel in a stand-alone claim, but that does not lead to the conclusion that time starts to run when the document becomes publicly available. In my judgement this approach is apparent from the analysis I have so far considered but is particularly clear from Neto, the final STJ case that I need to consider at this stage.

This was an appeal by SCL in proceedings commenced on 16 December 2019 by an orange producer Mr Neto concerning contracts for the forward sale of oranges entered into between 2002 and 2003. The purpose of the appeal was to decide the term of the limitation period for the claim and from what date it had started to run. The claim had been dismissed at first instance. That decision had been overturned by the Court of Appeal on the basis that the starting date of the limitation period should be the date of the Final Decision of CADE approving the TCCs. The STJ allowed the appeal from the Court of Appeal with four judges (led by Andrighi J as the Judge Rapporteur) deciding to allow the appeal and one judge (Justice Cueva) deciding to dismiss it.

Andrighi J started by summarising the relevant proceedings below:

“4.

On the one hand, the trial court, adopting the appellant's argument, issued a ruling stating that the limitation period began with the widespread dissemination of Operation Fanta in the media, starting on 24/1/2006 (e-STJ page 574). In this context, it was decided that “since the deadline for exercising the claimant's claim started on 24/01/2006 and the lawsuit was filed only on 16/12/2019, the limitation period should be rigorously recognised, based on Article 206, Paragraph 3, V, of the Civil Code” (e-STJ page 575).

5.

In the opposite direction, adopting the Defendant's thesis, the Court of Appeals of São Paulo, by majority, decided that the limitation period began on 6/3/2018, that is, with the final order of CADE that approved the TCC and ended the investigation pending before the administrative body.”

The judge then draw attention at paragraph 16 of her judgment to the distinction between stand-alone and follow-on claims by reference to earlier cases already considered and adding that a follow-on case is one in which “… the alleged violation of economic standards depends on a decision by CADE…” with all other claims for damages and other relief by reference to alleged anti-competitive conduct being stand-alone claims. She then held at paragraph 18 that in relation to follow-on claims as defined by her, the limitation period begins to run “… with the final sentence order issued by CADE…because it is from that moment onwards that the individual becomes unequivocally aware of the violation of the right and its extent, relying on the theory of actio nata in its subjective bias.” On the other hand, in relation to stand-alone claims, setting when the limitation period starts to run “… requires a case-by-case examination of the moment when the holder of the right became effectively aware of its violation, as there is no interference from the administrative body or, at least, there is no adverse decision recognising performance of the wrongful act.”

Against that background, the judge turned to the impact of a TCC on when the relevant limitation starts to run. The Judge defined such agreements as being agreements “… signed between CADE and companies or individuals investigated for alleged antitrust violation. Through this instrument, the signatories undertake to suspend the practices that generated suspicions of antitrust conduct and also undertake to pay financial contributions. In addition, other measures may be established to stimulate or reestablish competition in the market”. The judge then referred to Article 185 of the CADE Internal Regulation Law, which described a TCC as “… an instrument enforceable out of court,[which] must specify the obligations of the represented party not to engage in the conduct under investigation or its harmful effects, as well as any obligations that it deems applicable; setting the penalty amount in case of full or partial non-compliance of undertaken obligations…” and concluded (at paragraph 24) that:

“The Agreement, therefore, is established on a case-by-case basis, according to the peculiarities of each administrative proceedings and in light of ongoing investigations. In fact, it is possible that in some Agreements there is effective recognition of the practice of the tort by the committed party, while in others there will be no such admission.” [Emphasis supplied]

The judge then proceeded to consider in detail the effect of the CADE Final Decision and the TCC in terms which again I must substantially reproduce because it represents the fullest consideration by the highest relevant court in Brazil of these matters, which as I have said are central to the claimants’ case that the earlier decisions referred to above stemming from Teles are wrongly decided.

The judge rehearsed the findings of the lower court in detail at paragraphs 31-32, in the course of which she recorded that:

“32.

… the ruling recognised that the Cessation of Conduct Agreement was signed without acknowledging the illicit practice, verbatim:

“[T]his case differs in relation to the legal nature of CADE's decision. While the case law was based on the understanding that the limitation period starts from the decision that recognises the practice of the wrongful act (of a sentencing nature, therefore), in the situation of the case file, the decision was ratified by the Terms of Commitment for Termination - TCC.

The claimant claimed, in response to the statement of defence, that there was recognition of the defendant's conduct for the formation of the TCC, in accordance with Clause Two of the Terms of pages 83-90, which provides:

2.1.

In accordance with the requirements in the applicable legislation, the execution of this Term of Commitment implies the admission, by the COMMITTED PARTIES, of the facts described in the History of Conduct, which consists of an integral part of this term as Annex I below.

In Annex I of the executed Term, the first item contains the following:

The COMMITTED PARTIES recognise that information gathered by their commercial team in Brazil from the citrus market, especially with independent orange growers and traders, may have been occasionally shared with competitors in the context of sectoral discussions on this market, the same occurring with equivalent information obtained from the market by its competitors during the investigated period (page 354)...”

The judge records that in the course of the lower court’s judgment, the majority had held that:

“The recognition, although occasional, carried out by the defendant, dealt exclusively with information sharing. This sharing of information, without supplementary data, without the existence of a context signed in the TCC or, in any case, without the express statement that there was an assumption of fault about facts that constitute the core of the act defined as forming a cartel, cannot serve the purposes intended by the applicant.

Therefore, established that there is no final decision by CADE that recognises the practice of the fact stated in the particulars of the claim (formation of a cartel), the initial term of counting the limitation period is that established in Article 189 of the Civil Code (actio nata), with the addition of subjective bias (knowledge of wrongdoing, enforceability of the right), as well as in Article 47 of Law No. 12.529/2011…”.

The judge also records the conclusion of the dissenting judge in the Court below as being:

“"I would like to emphasise that unequivocal knowledge of the facts (Operation Fanta) should not be confused with unequivocal knowledge of the certainty of the fact (execution of the TCC). In fact, the claimant's claim is doomed to the most irremediable failure because it gives the TCC a certainty (conviction/confession) that did not occur.

In this vein, guided by brevity, it is essential to transcribe the response to the questions asked in the opinion:

'First question: We question whether the signing of the Term of Commitment for Cessation by SUCOCÍTRICO CUTRALE, in 2016, implied admission of the practice of a cartel, which served as a cause of action for the Claimant.

The answer is no: the signing of the Term of Commitment for Cessation by SUCOCÍTRICO CUTRALE, in 2016, did not imply admission of the practice of a cartel, which served as a cause of action for the Claimant. A "summary description of the conduct," as set out in the TCC, does not mean that the signatories practice price fixing or market division, observing, in the summarised data, only a potential and occasional sharing of information without this proving a combination.”

Andrighi J (who as I have said was the Judge Rapporteur for this case) then continued:

“33.

Based on the above, since the Cessation of Conduct Agreement (TCC) was approved without recognition of the practice of the violation, it is necessary to examine when the appellee became effectively aware of the violation of its right.

34.

Hence, the initial point for calculating the limitation period for the specific case addressed in the case file must be defined.

35.

Initially, this Rapporteur was inclined to follow what was established in the ruling, recognising the initial term of the limitation period as of the wide dissemination of news about the investigation of Operation Fanta in the media, starting on 24/01/2006.

36.

However, in the session of 18/02/2025, Justice Moura Ribeiro disagreed with the aforementioned conclusion, suggesting that the date of the agreement established between the farmer and the juice-purchasing industry be adopted as the initial term, given that at that time the price for payment of the oranges was set at an amount clearly lower than that practiced on the market. In other words, the appellee was aware of the situation and could have taken action against the appellant from the beginning of the contractual relationship.

37.

In fact, considering the instructive observations made, I adhere to the solution proposed by Justice Moura Ribeiro, in order to conclude that the beginning of the limitation period, with the unequivocal knowledge of the violation of the defendant's right, occurred from the date of the conclusion of the orange purchase and sale agreement, in the years 2002 and 2003.

38.

Consequently, since the initial term occurred between 2002 and 2003 and the claim was filed only on 16/12/2019…, the limitation period lapsed…

40.

Therefore, the state appellate decision must be overturned, thus reestablishing the ruling that extinguished the trial with prejudice due to the limitation period of the claimant's claim.”

Mr Neto applied for “clarification” from the STJ maintaining that amongst other things the STJ had failed to consider Article 185 of CADE’s Internal Regulations – which as I have explained the claimants in these proceedings rely on as negativing the reasoning in the STJ decisions to which I have so far referred. The application was dismissed by the STJ. Having rehearsed that the application concerned, as one of the two issues that required clarification, an omission regarding Article 185 of CADE’s Internal Regulations, the STJ rejected that as a ground for clarification because “… the judgment contains no omissions, as all provisions alleged to have been violated were duly examined by the Collegiate of this Third Panel. It should also be noted that it is widely accepted in this Court that a judge is not obliged to address all the arguments put forward by a party if there is a valid reason for not doing so…”.

Since the conclusion of the trial there have been a number of decisions of the STJ that follow or are consistent with the reasoning of these cases referred to above.

On 28 August 2025, the STJ published its decision in FS-A Special Appeal No. 2154646 - SP (2023/0272053-9) (“FS-A”). The Judge Rapporteur in that case was Justice Teixeira. The appeal was from a decision of the Court of Appeal holding that the three-year limitation period began with the publication of the CADE Final Decision that ratified the TCC and ended the administrative investigation in 2018. The appellants argued that the limitation period should be calculated from the date of the supply contracts (2002/2003) as the TCC did not imply recognition of illegality. The STJ summarised the law as being:

“3.

Case law from the Superior Court of Justice distinguishes between follow-on claims, where there is a condemnatory decision by CADE recognising the unlawful act, and stand-alone claims, where liability is discussed directly in court without prior recognition of the unlawful act by the administrative body.

4.

In follow-on claims, the limitation period begins with CADE's final condemnatory decision. For stand-alone claims, the deadline begins from the moment the claimant was unequivocally aware of the injury suffered, which requires a case-by-case examination.

5.

Ratification of a Cease and Desist Agreement (TCC) without acknowledgement of the unlawful practice does not constitute a condemnatory decision, nor does it suspend or interrupt the limitation period.” [Emphasis supplied]

The STJ allowed the appeal because:

“6.

The present case involves a stand-alone claim since CADE did not expressly recognise the cartel practice of the investigated companies, limiting itself to executing commitments with the provision of obligations and payment of a fine without admitting guilt.

7.

The initial term of the limitation period must be set on the date of execution of the contracts between the orange producers and industries between 2002 and 2003, when the damage resulting from artificially low prices allegedly occurred.

8.

However, the claim, filed in December 2019, exceeded the three-year period provided for in Article 206, para. 3, V, of the Civil Code and therefore the claim for damages is time-barred.” [Emphasis supplied]

The STJ observed that in reaching the contrary conclusion, “… the Court of origin departed from the interpretation of this Court of Justice. In cases involving identical issues to those under analysis, it was determined that the initial limitation period would commence on the date of the contract signed between the producer and the juice industry, as this would be the moment when the price was theoretically fixed at a level lower than the market price.” This analysis was adopted unanimously by the rest of the court.

Following completion of the hearing, a further decision was delivered by the STJ that was said to be material. Following my attention being drawn to this decision, I directed that there be short additional written submissions from the parties addressing this decision. Those submissions are contained in letters to me from the defendants’ solicitors dated 8 October 2025 and from the claimants’ solicitors dated 13 October 2025. In addition, by a letter to me of 15 October, the claimants’ solicitors informed me of another decision of the STJ, that they submitted assisted their client’s case. For reasons that I explain below, I do not consider that case assists on the issues that actually matter in this case.

The decision that was the subject of the 8 and 13 October submissions is Fabri v. Louis Dreyfus Company Brazil SA Special Appeal No. 2908997 - SP (2025/0130672-0) (“Fabri”). In that case, the Court of Appeal had concluded that the claim was statute barred because “… the decision that ended CADE's proceedings did not recognise the formation of the orange juice cartel. This does not represent the beginning of a limitation period for compensation for damages, since no unlawful act was recognised.” The claimant had alleged both to the court below and in the appeal to the STJ that Article 185 of CADE’s Internal Regulations had been ignored by the courts below in arriving at that conclusion. The STJ concluded that the appeal should be rejected for reasons including that “… case law from the Superior Tribunal of Justice (STJ) has established that, once CADE rules out the existence of a cartel, the limitation period for reparatory claims begins from the execution of the contract for the purchase and sale of oranges.” This conclusion was adopted unanimously by the panel of the STJ to which the case had been allocated.

I accept the defendants’ submission that this was another decision of the STJ in which a claim in relation to the alleged orange cartel is time-barred. In reaching that conclusion the STJ had necessarily rejected the submission that Article 185 was relevant to the issue or had any material effect on the proper interpretation of the CADE Final Decision or the TCC to which the CADE Final Decision gave effect by discontinuing the proceedings against the signatories to the TCC after they had complied with its requirements. Whilst I accept that this is another decision that reflects the settled position of the STJ, it adds nothing of substance to the reasoning adopted in the earlier decisions. It is as vulnerable as the others to the claimants’ submission (if it is correct) that all of these STJ cases have fallen into the same error of failing to appreciate that there has been a recognition by CADE of unlawful cartel conduct either by reason of paragraph 86 of the reasoned vote leading to the Final Decision of CADE and/or the implications that are said to arise from the effect of Article 185 of the CADE’s Internal Regulations.

I can deal much more quickly with the other more recent decision relied on by the claimants’ solicitors in their letter of 15 October 2025. This decision concerned an entirely different alleged cartel concerning the concrete industry. This decision - Votorantim Cimentos SA V. Concremix SA Special Appeal No. 1567390 - SP (2019/0245300-5) (“Concremix”) - does not assist because there was no dispute in that case that the courts were concerned with a follow-on case – see paragraph 1 of the judgment. This reflects what was said in Ramos, where it was held in relation to the cement cartel, that CADE had recognised the existence of the cartel in its final condemnatory decision on 22 September 2015. Concremix and Ramos show what must be obvious – that not all final decisions are condemnatory decisions in the sense of making findings concerning the existence of an alleged cartel that can then be used by a claimant as the basis for claiming damages. That is the reasoning that lies behind each of the decisions by the STJ referred to above.

More generally, Concremix was concerned with an argument concerning the effect of a claimant knowing the relevant facts long before the publication of the administrative decision that founds the follow-on claim being advanced. The proposition that in such circumstances, time started to run before final publication of the CADE decision relied on as supporting a follow-on claim was rejected because “… the STJ holds that, in follow-on compensation claims, the limitation period applies to the claim as a whole and begins on the date CADE's condemnatory decision is published. If the claim is filed within this period, full compensation for damage is permitted, including for those that occurred before the three-year period preceding the filing of the claim.” This is immaterial to this case because here the issue is whether the claims made by the claimants can properly be treated as follow-on claims given the terms of the TCC and the final determination of CADE.

The defendants maintain that on proper analysis this is not a follow-on claim at all – in essence because (as the STJ has held in the cases referred to above) – the TCC does not contain any relevant admissions and it and the Final Decision by CADE does not contain any findings on which liability could be based; that in consequence each of the claims is on proper analysis a stand-alone claim and applying the actio nata principle summarised earlier, time began to run when the claimant had “unequivocal” knowledge of the alleged violation of his right. For these purposes knowledge is not only what an individual wrong-doer actually knows but also knowledge such victims are presumed to have because it was discoverable with reasonable diligence.

The defendants submit that all the claimants either knew or could reasonably have acquired all the relevant knowledge required by the later of either the reporting of Operation Fanta and the publication of the notice concerning CADE’s administrative proceedings in January to February 2006 or (if later) when the claimants were compelled by the alleged cartel to enter into the forward sale agreements on uneconomic terms with the processor that is said to have inflicted the losses for which damages are claimed. The claimants maintain that the claims are all follow-on claims and as such time started to run only with the execution of the TCC and CADE’s Final Decision. To the extent that the decisions set out above suggest a different outcome they are wrong and result from an error first made by the STJ in Teles.

As I emphasised at the start of this judgment, the function of the English court in a case such as this is to predict the likely decision of a foreign court on the issue that arises and, once it has determined what the likely decision will be, to give effect to it in the English proceedings unless it is satisfied on the whole of the evidence that such a decision does not accurately represent the law having given full weight to the decisions of the most senior court in the jurisdiction concerned.

I accept that the rigid stare decisis approach that applies in common law jurisdictions does not apply in the same way in civil law jurisdictions. However, the degree to which respect is accorded to previous decisions of the highest relevant court in any particular civil law jurisdiction will vary. As I indicated earlier, the function of the STJ is ensure uniformity of interpretation of non-constitutional federal legislation, which is the source of Brazil’s limitation law. This means that it is probable that all the lower relevant courts in Brazil will follow the decisions of the STJ. That this is probably so is apparent from the terms of the later judgments in the sequence of cases decided by the STJ referred to above. Aside from two regional court of appeal decisions (overturned on appeal to the STJ), all the lower courts have all followed the decisions of the STJ in Teles and the cases that followed. This shows that lower courts in Brazil recognise the role of the STJ in ensuring uniformity of interpretation of non-constitutional federal legislation. Thus whilst I accept that if it could be demonstrated that Teles and the cases that followed were wrongly decided they should not be followed by an English Court applying Brazilian law, this will be a high hurdle to overcome given the role of the STJ in Brazil’s legal system and the notably consistent approach that has been adopted by the STJ ( and the lower courts) to this issue in the cases it has decided.

The key issue that lies at the heart of this debate is whether CADE’s Final Decision was “condemnatory” and/or the relevant TCC contains admissions of wrong-doing sufficient to enable a claimant to proceed to launch a follow-on claim. The advantage for a claimant of having a follow-on claim available is that the claimant is able to use the findings made by the relevant regulator as establishing liability leaving the claimant to prove the losses it or they allege to have been caused by the anti-competitive conduct established by the regulator concerned. This reflects evident legal (indeed common) sense since otherwise there would be no obvious reason for distinguishing between stand-alone and follow-on claims other than as a means of circumventing the limitation principles that apply to stand-alone claims. That would be unprincipled, whereas an approach which enables a claimant to avoid the cost and uncertainty of attempting to prove liability in a competition case by relying on findings of anti-competitive activity by a regulator and having only then to prove causation and loss is principled.

I reject Professor Frazão’s evidence to contrary effect as being contrary to what the STJ has held in at least Teles, Fabbri and Neto and because there is no principled reason for such an approach. If such was the case then there would no need for any test other than one based on time running from the moment of the victim acquiring (or being reasonably able to acquire) knowledge sufficient to start time running for limitation purposes with it being a question of fact in each case whether time started to run earlier than the date of the publication of a relevant condemnatory decision. The approach I favour is not merely that adopted by the STJ in the numerous cases it has decided referred to above and is jurisprudentially principled but in fact reflects the relevant academic writing. The judge in Teles justified this approach identified by reference to the writing of Mr Bruno Oliveira Maggi who refers to the substantive difference between the two types of claim in terms that will be familiar to English (and EU) competition law practitioners in these terms:

“…claims seeking damages for competition damage may be based on anti-competitive conduct reported directly by the victims or conduct that was investigated by antitrust authorities. He explains that, in the first case, there is a standalone lawsuit, in which the victim presents the evidence of the alleged act, as well as the damage suffered. The second case is that of the follow-on claim, “in which the victim supports the entire claim on the evidence and decisions produced by the authority that judged and condemned the cartel” (Idem).

Maggi further states that the follow-on claims, in which the injured parties will only know that they are suffering damage when the cartel is revealed by the competent administrative authority, no chronological point that precedes CADE’s final decision could be considered as the date of the damage- event, nor would it be possible to determine its eventual consequences (damage-loss). In these cases, in which the action is based on a decision by the administrative agent, “it is clear that the milestone that determines the beginning of the limitation period is the verification of the damage-loss, which may or may not coincide with the damage- event and with the illicit act” (Idem).

Once the purpose of a follow-on claim being available is understood, the approach of having two limitation regimes then becomes understandable and principled. Where a claimant seeks to rely on the liability findings of a regulator, logic suggests that limitation cannot start to run until the regulator has published its final findings because it is only when the findings are published that the claimant can know the liability finding available on which to base its claim. That is why Concremix was decided as it was. It is an approach which receives clear support from the judge rapporteur’s conclusion in Neto that a follow-on case is one in which “… the alleged violation of economic standards depends on a decision by CADE…” and appears also to have been the point being made by the judge rapporteur in FS-A. However, where there are no findings made by the regulator, it is difficult to see how there can be a follow-on claim because there are no liability findings on which the follow-on claim can be based for the purpose of establishing either causation or loss. On this basis, unless the claim is a follow-on claim in which the claimant is able to rely on the liability findings made by the regulator concerned (in this case CADE) then by definition the claim is a stand-alone claim in which the claimants must prove every aspect of the liability case they seek to advance.

Whilst the limitation response will vary from jurisdiction to jurisdiction, the response of Brazil’s legal system is to treat the limitation period for both stand- alone and follow-on claims as depending on the application of the actio nata principle to mitigate the literal effect of Article 206(3)(v) of the BCC by treating time as starting to run for a follow-on claim on the date when the decision being relied on to establish liability became available and for all other (inevitably stand-alone) claims, when the claimant had or reasonably ought to have had unequivocal knowledge of the (or a possible) wrong, that it had an adverse (or possibly adverse) effect on that claimant and the (or the possible) identity of the alleged wrong-doer. It is a consequence of this approach that limitation might start to run for a follow-on claim even though the limitation period for a stand-alone claim has been lost – an example of this being Concremix.

In the case law referred to above, the STJ has consistently rejected the submission that CADE’s Final Decision was condemnatory. It is not necessary that I go through each of the cases that I have referred to in detail already. The claimants’ key point is that the STJ in reaching this conclusion has misunderstood the TCC and the Final Decision of CADE largely because the court has misunderstood the terms of those documents or failed to construe them in their context which involves an examination of what the claimants contend is required as a matter of law (including specifically the laws applying to the conduct of CADE’s activities) before a TCC can be entered into or an investigation suspended by reference to such an agreement. The claimants submit that this error led the STJ to characterise the claim in Teles as a stand-alone claim when it should have characterised the claim in that case as a follow-on claim and that the STJ has thereafter consistently failed to correct the error, even though what is alleged is the error has been drawn to its attention.

In advancing this argument, the claimants submit that CADE is authorised to enter into TCCs only to the extent permitted by its Internal Regulations and CADE’s Internal Regulations (referred to by the experts in these proceedings by the acronym RICADE) were amended with effect from 2013 so as to expressly require by Article 185 an admission of participation in the conduct under investigation as a condition for entering into a TCC.

Article 185 provides that:

"In regard to investigations into agreements, arrangements, manipulation, or alliances amongst competitors, Cease and Desist Agreements must necessarily contain a statement by all signatories admitting their participation in the conduct under investigation."

In my judgement this provision is not one by which I can conclude that the STJ has wrongly decided the cases to which I have referred above. Even if it was to be accepted that CADE acted in breach of Article 185 by entering into a TCC with an alleged cartelist that did not contain “… a statement by all signatories admitting their participation in the conduct under investigation…” that might lead to the conclusion that CADE had acted unlawfully in entering into the relevant TCC but that does not lead to the conclusion even arguably that a TCC that does not contain such a finding or admission should be construed nonetheless as containing one. The STJ in both Neto and Fabbri distinguished between an admission of an illegal act or an admission of a part of the conduct under investigation. As I explain in more detail below, TCCs do not confer immunity from criminal prosecution and are typically heavily negotiated between lawyers acting for parties under investigation and CADE. This is likely to explain why what is conceded in them by those being investigated is likely to be very limited and highly qualified.

In the TCC relevant to this case what is admitted is both very limited and highly qualified. Although the claimants maintain that what is admitted is by its very nature unlawful, this is an argument that has been repeatedly rejected by the STJ in Teles and in every case since concerning the alleged cartel that is the subject of these proceedings. The rationale for that conclusion is set out in the court’s judgment in that case – it is that (a) a commitment to cease the practice under investigation “… does not imply a confession as a matter of fact or recognition of the unlawfulness of the analysed conduct…” and (b):

“…the recognition referred to by the Claimant, in fact, is that there may have been, possibly, the sharing of information. There is no recognition, by the Defendant, of the practice of the acts indicated by the Claimant in the opening brief, not even a decision by CADE recognizing the existence of a cartel, or confession by the Defendant in relation to this fact. Thus, it is necessary to conclude that the recognition, even if occasional, carried out by the Defendant, referred exclusively to the sharing of information. This sharing of information, without additional data, without the existence of a context established in the TCC or, in any event, without the express statement that there was an assumption of guilt regarding facts that constitute the core of the act defined as cartel composition, cannot serve the purpose intended by the Claimant.”

The claimants further submit that in a cartel case, an admission of participation in the investigated conduct should be treated as an admission of participation in the cartel alleged. I do not accept that is a correct analysis but nor do I accept that what the claimants allege is the policy requirement for such an approach has been made out either.

The underlying public policy rationale for this approach is said to be that it encourages parties to take advantage of the availability of leniency agreements to confess and cooperate in exchange for immunity from administrative and criminal sanctions by making the options left for someone who has not sought leniency less attractive. This is said to lead to the conclusion that very limited and highly qualified admissions should nonetheless be construed as matter of law in far wider terms that they are expressed or intended to be read.

Professor Frazão maintains that in policy terms it would make no sense to require a confession from someone seeking leniency but not from someone seeking to enter into a TCC. As she puts it in paragraph 148 of her second opinion:

“In effect, the requirement for explicit confession in TCCs seeks to preserve incentives for leniency to remain the most advantageous option for those who collaborate from the outset of the investigations. Therefore, this calibration is essential to the effectiveness of Brazilian antitrust policy, ensuring that each instrument fulfils its role and promotes a balanced competitive environment.”

She goes so far as to say that:

“Hence, from the perspective of Competition Law, there is no other possible interpretation other than that of the impossibility to enter into TCCs in cartel cases without admission of participation in the practice under investigation, which is equivalent to confessing the unlawful nature of the conduct,”

and that “… the admission of participation in the unlawful conduct under investigation is a sine qua non condition for the execution of TCCs in cartel cases.”

I do not accept this evidence. Firstly, at no stage has the STJ ever acknowledged that this is the correct approach to be adopted when considering the true meaning and effect of a TCC agreement. Secondly, I do not accept the logic that is said to support this approach. The usual benefit to be obtained from a leniency agreement in a competition law context is that it protects the applying party from or mitigates the degree to which that party is exposed to, the very substantial penalties that apply to those who do not enter into leniency agreements including by conferring immunity from criminal prosecution. Generally, leniency will be available only to the first participant who applies for such a benefit. The remaining parties must either settle through TCCs or wait for the regulator's decision following the completion of its investigation. The benefit of entering into a compromise such as a TCC is that it mitigates rather than eliminates the risk of penalties which can be eliminated only by entering into a leniency agreement. It does not however confer immunity from criminal prosecution.

Whilst I accept that the contrary argument was advanced in later cases, it is worthwhile noting that in the judgment in Teles express reference is made to academic writing on this point by Flávio da Silva Andrade in “Reflexões sobre os instrumentos de consenso na defesa da concorrência: acordo de leniência e termo de compromisso de cessação de prática antitruste.” Revista CEJ, Brasília, Ano XXI, n. 71, Jan./Apr. 2017, pages 109- 121 which supports this analysis rather than that of Professor Frazão. In the section of his work quoted in Teles, Mr da Silva Andrade states:

“The Cease and Desist Agreement (TCC) is an instrument that allows the solution of a conflict based on the negotiated application of the norm. There is the possibility of the applicant participating in the construction of a harmonic solution for the case. Thus, since the main objective is to commit to ceasing the investigated harmful conduct, one should not demand the confession of an infraction or the acknowledgment of participation in the investigated conduct by the applicant, since, as seen, the applicant is not assured criminal immunity.”

The Judge also quotes from another academic writer (Geisa de Assis Rodrigues) in the following terms:

“Thus, conducting the adjustment must be carried out in a non-repressive perspective, without pre-conceived ideas. Therefore, there does not need to be an explicit acknowledgment of guilt on the part of the obligee, since in many cases this is absolutely irrelevant. What is really wanted is the cessation of the conduct that offends the trans-individual right. The lack of need for confession or acknowledgment of guilt is an element that facilitates the conclusion of the adjustment, which is often not achieved in court”.

Thus, whilst I accept that Article 185 RICADE is in the terms set out above, I do not accept that it has the effect described by Professor Frazão. That is not its effect on its face, and it is not an effect that would make any sense in policy terms, which makes it unlikely it would be construed as having the effect contended for. In any event the TCC in this case does not have the effect alleged as a matter of construction applying the applicable principles of construction, which are Brazilian law principles that have been applied consistently by the highest relevant Brazilian court to the TCC in this case. Finally even if this is wrong and what appears in the TCC is construed as being an admission by the investigated parties to a TCC of “… their participation in the conduct under investigation…” it begs the question whether what has been admitted is capable of establishing liability in a private law claim between a producer and a processor operating within the alleged cartel because it is only if that is so that a claim can be said to be a follow-on claim. If what has been admitted cannot form a platform for a claimant to prove causation and loss there is no follow-on claim available. In my judgement what has been admitted in the TCCs does not come close to having this effect as the STJ has held consistently. I expand on this conclusion later in this judgment.

The claimants place significant weight on the fact that CADE applied to intervene and file amicus briefs in Fabbri and Neto. The point that the claimants make is that if that had been permitted then the policy points to which I have referred would have been fully deployed and the STJ would have arrived at a different conclusion concerning the true meaning and effect of the TCC and/or the effect of the CADE Final Decision. I reject that submission for the following reasons.

Firstly, the points lack merit for the reasons already considered. Secondly, the points that arise were set out comprehensively in an opinion filed in Neto by the Brazilian Rural Society (“BRS”) as Professor Frazão accepted in her oral evidence:

“MR KENNELLY: Professor, we were looking at subparagraph (iv), which you had before we broke, and this is a reference to the new Article 185 introduced to the CADE internal rules in 2013; correct?

A. Yes.

Q. So that point was before the STJ in Neto, wasn't it?

A. Yes.

…

MR KENNELLY: Professor Ragazzo made to the STJ [the point] that TCCs should be construed so as not to undermine the leniency programme.

A. Yes.”

The claimants argue that although the point was placed before the STJ, it failed to grapple with the point. The claimants rely on the fact that motions for clarification were filed in Neto in which it was argued that the STJ had ignored important arguments they had raised. As I recorded earlier those challenges failed. Professor Frazão readily accepted that the effect of this process was that the arguments referred to were all presented to the STJ for a second time – see T3/745/1. The point she made was that “… none of the grounds that we are discussing here, including Article 185, were taken into consideration…” Professor Frazão accepted that the Article 185 point had also been before the STJ in Oliveira and (probably) Ramos. It was certainly considered by the Court of Appeal in Fabri - see the summary set out earlier in this judgment. The STJ dismissed the appeal from the decision of the Court of Appeal in that case.

I accept that the issue was not addressed expressly by the STJ in its judgments. Whilst the practices of courts differ from jurisdiction to jurisdiction, the reasoning adopted by the STJ in Brazil appears to be much more concisely expressed than the decisions of courts and particularly ultimate courts of appeal in other jurisdictions. However, the STJ set out in its judgments when considering review applications, the well-established Brazilian law procedural principle that the court was not expected to comment on all points deployed before it. In my judgement significant caution should be exercised by an English court before concluding that a foreign court had decided an issue wrongly by reference to the omission from a judgment of an express ruling on the point, when that court expressly states that is considers that all relevant provisions had been examined. In my judgement it is entirely unreal to suppose that the STJ would decide the issues that arise any differently from how they have been decided to date by reference to the point I am now considering.

In my judgement the proposition that the STJ would decide the issue I am now considering differently if asked to consider it again is all the more improbable because, as Professor Frazão accepted, the policy point summarised above concerning encouraging applications for leniency was set out in the BRS Brief – see T3/47/6. Professor Frazão made the same point a little later in her oral evidence where she also accepted that these points had been addressed in more detail than in the CADE amicus brief that had not been admitted:

“A … this argument is both in the CADE's opinion and both in Professor Ragazzo's opinion as well.

Q. It's actually set out in far more detail in Professor Ragazzo's opinion, isn't it, this argument that the leniency programme would be undermined if the TCC was treated as not amounting to an admission to infractions?

A. Yes. He explored this argument in more detail in his report.”

This is significant because the author of the BRS Opinion is Professor Ragazzo, a former General Superintendent of CADE.

It was suggested to Professor Didier during his cross examination that the absence of an amicus brief from CADE itself was significant because of the role that CADE plays in enforcing competition law in Brazil. In my judgement there is no substance in that point. Whilst it is likely that a court would show some deference to a specialist body where the issue in question was a technical one falling within the specialist expertise of the institution concerned, that is much less likely to be so where the issue is one of law and in any event this point was not one that Professor Frazão considered significant: as she put it “… the duty of deference is not an important fundament…” In fact in at least one decision of the Supreme Federal Tribunal (Brazil’s ultimate constitutional court) in Cascol Combustíveis Para Veículos Ltda and others v. CADE that court concluded that whilst a court had control over issues concerning competence, purpose, form, object and motivation, it was not entitled to substitute its own evaluation of economic effects when assessing a sanction for those were within the expertise of CADE as the regulatory and supervisory body of economic activity to which the legislature has assigned such decision making. This approach reflects the position in many jurisdictions in relation to specialist tribunals and regulators (including in England and Wales) and is one that I accept would be applied by Brazilian courts at all levels.

It is wrong to suggest that all the subsequent cases are wrongly decided because they followed Teles and ignored the arguments concerning Article 185 of RICADE and/or paragraph 86 of the reasoned vote leading to the CADE Final Decision. This is because those points were addressed in later cases before the STJ, especially in Neto and Fabbri. Most recently in FS-A, those cases were followed with the Judge Rapporteur noting that the court below had departed from the reasoning of the STJ which was then followed by all the justices including the Judge who had dissented in Neto and Fabbri. The argument is one that was fully deployed before the STJ in the ways that I have described. It is one that depends ultimately on the content of the TCC.

The CADE amicus brief accepted that “... there was no decision in a sanctioning administrative proceeding before [CADE] identifying the authorship and materiality of the unlawful competition…” (as Professor Frazão accepted during her cross examination see T2/124/1-3). and the relevant TCC does not contain any admission of participation in the alleged cartel. What was admitted was what was set out in Annex 1 to the TCC under the heading “History of the Conduct” – see clause 2.2 of the TCC.

As noted earlier, the only conduct referred to in Section I of Annex 1 was that the relevant parties acknowledged only that “… information gathered by their commercial teams in Brazil from the citrus market, especially from independent orange producers and traders, may have been possibly shared with competitors in the context of sectoral discussions on this market; the same may have occurred with equivalent information obtained from the market by its competitors during the period under investigation.” Section III in Annex 1 refers only that the relevant parties acknowledged that the identified entities “… may have participated in the acts mentioned in this Conduct History…”. Section IV acknowledges only that the conduct set out in Section I “… may have taken place during industry meetings at the Brazilian Association of Citrus Exporters…” Annex 1 concluded with Section VIII, which records only that the relevant parties came before CADE’s Chief Superintendent Officer “…to acknowledge the conduct described above…”.

The highly qualified terms in which the TCC is couched is unsurprising. Whereas leniency agreements (such as those entered by Mr Paulo Ricardo Soares da Cunha Machado, referred to earlier) confer immunity from criminal prosecution, TCCs do not. Whilst leniency agreements require an admission to unlawful conduct, what appears in TCCs is heavily negotiated between lawyers acting for the relevant parties and CADE, with the lawyers acting for the parties entering into the TCC being anxious to protect their clients from the risk of criminal prosecution. Professor Didier described a leniency agreement as having “… a characteristic which is quite peculiar. It is the fact that it's an agreement with CADE whereby it is demanded the recognition of the unlawfulness of the behaviour, and it is precisely because of that that the beneficiary of the lenience has immunity from prosecution…” and when he was asked about this in cross examination, the relevant exchanges were:

“Q. Professor, I'm wholeheartedly agreeing with you. My point to you is that it is sufficient within the TCC merely to confess to facts. The legal characterisation of those facts is then a matter for CADE and, as we have seen, CADE has decided that those facts amount to a proof of cartel behaviour. Do you not accept that?

A. I don't accept it, precisely because TCC is a commitment from CADE to not investigate this conduct that's been recognised, when it is recognised. So CADE is saying that if the person entering into the TCC recognises the conduct, CADE will not investigate or punish them for that. So I will say again: when CADE agrees to have the agreement with the recognition of the conduct and the payment, CADE waives the right to commence proceedings and to assess and get into findings with regard to illicitness. This is part of the agreement with CADE, an agreement that was concluded and that was complied with, and it ended, and it's existed for so long now.”

Professor Frazão agreed that a leniency agreement, unlike a TCC, gives immunity from criminal prosecution to the party to it – see T2.100/16-20, where she also agreed that a party to a leniency agreement “…must confess to having participated in the tort…” Professor Frazão considered there was no difference between the requirement for confessing to a tort in a leniency agreement and the admissions required for a TCC but in my judgement that is incorrect for the reasons already identified. In summary if that was so then the incentives to enter a TCC would be compromised and such an approach would fail to accord recognition to the essential difference between a leniency agreement which confers immunity from prosecution and a TCC, which does not. In answers to me, Professor Frazão agreed with the substance of the evidence of Professor Didier set out above:

“JUDGE PELLING: If you enter into a leniency agreement, then you are exempt from prosecution; is that correct?

A. Yes, that's the idea, my Lord.

JUDGE PELLING: All right. Now, with a TCC, that is not the case, is it?

A. No.

JUDGE PELLING: Much earlier in your evidence, you described the negotiations which lead to TCCs. Is that something which is handled, at any rate amongst sophisticated parties, by lawyers on the one hand for CADE and lawyers on the other hand for the alleged cartelist?

A. Yes.

JUDGE PELLING: This is as open as I can make it, really: would a concern in such negotiations from the cartelist point of view be to ensure that any admissions made are so confined as not to be capable of being used by prosecutorial authorities?

A.

Yes, this is a concern. This is the usual concern…”.

I accept of course that there will be cases in which CADE issues a condemnatory decision but not every decision it issues is condemnatory in the relevant sense of containing factual findings from which it is concluded that the parties named have been guilty of any competitive activities. This is something I have considered already above at some length. It is illustrated by the difference between the CADE Final Decision in respect of the cement cartel referred to earlier. It was also the point made by Professor Didier in the course of his cross examination by reference to the Industrial Gases case – see T1/138/17-25. The reason why the Industrial Gases decision is of no assistance is that identified by Professor Didier a little later in his cross examination, where he said that the “… decision on Industrial Gases is a case that presupposes a condemnatory decision from CADE. So it's slightly different. There was a condemnatory decision by CADE…” As he added when pressed on this point:

“I accept that Justice Cueva [in his judgment in Industrial Gases] said that. I agree with this view by Justice Cueva with the -- and I would like to highlight that it is only in the TCC where you have recognition of unlawfulness. But we must consider one thing. The case with the Industrial Gases is a case that involved the condemnatory decision from CADE. So as a precedent, it only has to do with the condemnatory decision by CADE. It's a theoretical consideration that Justice Cueva puts in his opinion, which I do agree with, but the case only involved a condemnatory decision from CADE. So there has to be a decision in a case where there's been a TCC recognising an unlawful act where the STJ compares or, even better, equates that, but in fact the STJ points out that in a different case which could come with a TCC where there was a recognition of unlawfulness, it would be treated in the same way.”

Turning to this case, CADE closed the case against the defendants at its 118th Ordinary Session of Judgment held on 28 February 2018, “… in consideration of their full compliance with the obligations established in the respective undertakings signed with CADE….”. One of the documents relied on by the claimants as part of the CADE Final Decision is described as the “Reasoned Vote as endorsed in the CADE Administrative Tribunal Decision” (“Reasoned Vote”).

Whilst it is true to say that paragraph 86 of the Reasoned Vote records on the basis of what is contained in the TCC that the relevant CADE Commissioner understood “… that the existence of collusive conduct has been proven. In the case of a cartel, that is, an offense per object, proven the materiality of the conduct, it is not necessary to analyse additional elements such as the effects, since the harmful potentiality is presumed of the anticompetitive object itself, as already stated by the CADE’s Tribunal …” what in fact was concluded by the vote was that the proceedings against the defendants in these proceedings would be ended due to their compliance with the TCC they entered into. What the commissioner said he understood is not either a condemnatory finding or an admission by the signatories to the TCC. It is for this reason that I conclude Professor Didier’s evidence that paragraph 86 is irrelevant should be accepted. The exclusive purpose of the CADE Final Decision documentation was to close the investigation without condemnation. Construing paragraph 86 as rendering the decision condemnatory is inconsistent with that purpose.

I accept the submission made by the defendants that the TCC, the decision at the 118^th session and the Reasoned Vote do not constitute either of themselves or together an admission of conduct to enable a follow-on claim to be made. That is so in their own terms. As I have explained the case was closed on the basis that the defendants had complied with the TCC. It does not anywhere contain a condemnatory conclusion to the effect that CADE has found the defendants to have breached any relevant provision of Brazilian competition law. The TCC itself contains no such admissions. Article 185 of RICADE does not lead me to alter these conclusions for the reasons that I have explained earlier. More importantly this material has been deployed before the STJ, and it has concluded that it has no effect on the limitation issue for that reason. It did so notwithstanding comprehensive submissions in the BRS Amicus Brief prepared by Professor Ragazzo. As with the TCC, the correct interpretation of this document is a question of Brazilian law. On each occasion when this issue has been considered by Brazil’s highest relevant court, it has concluded not only that the TCC contains no relevant admissions, but the CADE Final Decision is not condemnatory in any relevant sense.

The TCC is ultimately the only document that matters. Its true meaning and effect is a matter of Brazilian law. It has been considered multiple times by the STJ and on each occasion the same conclusion has been reached either unanimously or with one dissent. On each occasion, it has been concluded that that the TCC did not amount to an admission of participation in the alleged cartel. In arriving at that conclusion, the STJ on each occasion has examined clause 2.2 of the TCC and the parts of the History of the Conduct set out above. By the time the most recent case to come before the STJ prior to the end of the trial (MH) came to be decided the Judge Rapporteur in that case was able to say that “(i)t should be noted that the Superior Tribunal of Justice (STJ) has already ruled out the possibility of the TCC signed with the Brazilian Competition Authority (CADE) being evidence of cartel formation by the companies’ buying oranges. The local court could not have relied solely on this document to demonstrate that the respondent party had committed an unlawful act, as if it were equivalent to a confession of the facts alleged by the claimants/appellants…” Professor Frazão accepted that this as the effect of both that decision and the prior to decisions referred to in the judgment in that case – see T2/81/1-82/24.

For these reasons, I conclude that it is highly improbable that if faced with this issue again the STJ would diverge from the conclusions that it has consistently reached on this issue. This is apparent from the way in which the reasoning of the court has developed over time, that it has maintained the position in the face of submissions to the effect deployed by the claimants in this case and from the decisions reached in the most recent judgments of the Court. It has held consistently and in my judgement it is probable that it will continue to hold that the TCC does not contain any relevant admission and that the CADE Final Decision does not assist because it depends exclusively on the content of the TCC and is not condemnatory not least because it decides only that the case against the defendants be shelved because they have complied with their obligations under the TCC.

Before leaving this part of the claim, I should mention the publication of a CADE technical note in relation to an investigation by it into an alleged soybean purchasing cartel. This was not the subject of any evidence from either expert because it was published after the completion of the trial but before the final submissions. The claimants rely on this document because the document refers to the alleged orange cartel as a cartel that had been recognised by CADE. In my judgement it is questionable whether I should be considering a document of this sort without it having been subject to expert evidence. That said the claimants rely on it firstly because it demonstrates that illegality is to be assumed in relation to at least some anti-competitive conduct (paragraph 70) in relation to the alleged orange cartel an assertion in the document that “… Brazil has also had experience in recognising the anti-competitive illegality of purchasing cartels. Of particular example is the orange purchasing cartel, relating to Administrative Proceeding No. 08700.000729/2016-76. In this case, Cade investigated the actions of concentrated orange juice processing companies for having organised themselves to fix the prices to be paid for the inputs of their activities, as well as to divide the suppliers among themselves for this purpose. The unlawfulness of the conduct was recognised by this Authority, and the Respondents in this case signed a TCC…”.

The claimants submit that this document shows that “… CADE has always considered itself to have recognised the existence of a cartel in the 2018 decision. Further, the TCC entered into by the orange juice purchasers was entirely consistent with that finding of cartel practice.” In my judgement this submission is one I should reject. This document is not a decision by CADE as paragraph 145 of the document makes clear. The conclusion is that “… it is suggested that an Administrative Proceeding be instituted to impose administrative sanctions due to antitrust violations…” Paragraph 146 goes on to set out suggestions for the notification of the respondents so that they can present a defence. The document is in effect the case for the regulator against various identified respondents that has yet to be adjudicated on. Assertions in such a document are entirely immaterial to the issues that arise in this case. In any event, what is asserted is inconsistent with the terms of the Final Decision and the TCC that are relevant to these proceedings. What is asserted in a document generated in relation to another enquiry years after the events relevant to these proceedings is not material to the issues that arise in this case.

The claimants’ alternative case is that even if they fail on their case that their claims are follow-on claims with time starting to run on the date when CADE’s Final Decision was published, their claims are nonetheless not time barred because they could not have had sufficient knowledge to enable time to start running until after publication of CADE’s Final Decision.

As will be apparent from the summaries of the cases decided by the STJ referred to above, that court has consistently held that as a matter of Brazilian law, in relation to stand-alone claims time starts to run from the time when either a claimant had all the relevant knowledge necessary to enable that claimant to commence proceedings or could reasonably have acquired such knowledge. It is recognised by all parties in these proceedings that it is not practical to engage in an enquiry as to the actual state of knowledge of each claimant. As the claimants noted in their written opening submissions, “(a)n unusual feature of the trial is that the state of knowledge of the Cs is being investigated without directly hearing from any of them. This is inevitable, however. There are more than 1,400 Cs (22 of them companies and one a foundation). An investigation of their actual, subjective state of knowledge over time would have been unworkable.” The focus of attention therefore has been exclusively on what knowledge would reasonably have been available to the claimants and when. The defendants’ case is that the claimants all had or reasonably could have had the knowledge required to start limitation running by no later than the date when Operation Fanta was executed and the publication of an administrative announcement concerning the scope of the administrative proceedings that followed. The factual detail surrounding those events is set out earlier in this judgment.

Although the claimants submit that the issue I am now considering is not one that can be resolved by referring to prior decisions of the Brazilian courts I cannot agree. I accept of course the long-established principle that findings of fact made in earlier court decisions not involving the same parties are inadmissible applying the decision of the Court of Appeal in Hollington v. Hewthorn [1943] KB 587. However, the Brazilian courts generally and the STJ in particular has consistently held as a matter of law that claimants seeking damages for losses caused by the alleged orange cartel had or could reasonably have acquired the knowledge necessary to enable time to start running for limitation purposes at the later of either the date when Operation Fanta became public knowledge and the publication of an administrative announcement concerning the scope of the administrative proceedings or the date when the relevant supply contracts said to have caused the relevant losses were entered into by the relevant claimant with the alleged cartelist concerned. That is a conclusion which is a mixed question of law and fact. It is a conclusion of law to the extent that it is a conclusion that those events were a sufficient basis for concluding that a claimant could reasonably have acquired the necessary knowledge to start time running for limitation purposes. The facts themselves are not in dispute. In relation to that question therefore, the issue that an English court must decide is whether that conclusion would be reached by the Brazilian courts if the issue came before it again.

The STJ has held consistently that the limitation period starts to run from the moment the victim becomes aware of the unlawful act and the identity of the wrong-doer – see Teles at paragraph 6. In arriving at that test the STJ has attempted to balance the absurdity that would arise from enforcing Art. 206 of the BCC literally, with the need to ensure that limitation continues to run where there has been “effective inertia” on the part of the victim – that is a failure to take action notwithstanding that the victim knew or ought reasonably to have known of the unlawful act and the identity of the wrong-doer. Having concluded that a follow-on claim was not available for the reasons considered already, the court in Teles then concluded that time started to run from the date when the claimant entered the relevant contracts with the alleged cartelists, which had been signed in 2001 and 2003. The point made by the court below and adopted by the STJ was that the claimant in that case was aware of the relevant facts at the date when the contracts were entered into.

In Ramos the STJ upheld the conclusions reached by all the lower courts that time started to run on 24 February 2006 with the publication of the initiation of the administrative proceedings because that was when the “(u)nambiguous knowledge of the possibility of an infringement…” occurred because thereafter “… the interested party could take legal action to preserve or assert rights.” Having rejected the suggestion that the claim should be treated as a follow-on claim for the reasons set out above, the judge continued:

“It is important to consider the point at which the potentially injured party became clearly aware of the alleged existence of the cartel and the possibility of having suffered damage, and could then take legal action to preserve or enforce its rights, in which case it must prove not only the causal link and the damage, but also the conduct alleged to have caused the damage.

In the present case, the initiation of administrative proceedings at CADE to investigate the formation of a cartel (art. 20, items I, II and IV with art. 21, items I, II, III, IV, V, VI, X and XII of Law 8.884/94) against the appellee companies and others was published in the Official Gazette of 24 February 2006 (page 284), at which point the injured parties could take the legal measures they considered appropriate to protect their rights or to claim civil damages.

Therefore, I believe that the three-year limitation period for non-contractual civil damages (Art. 206, para 3, V, CC) should be adopted as the starting point, the date on which the opening of administrative proceedings at CADE against the appellee companies was published on 24 February 2006, the date from which the potential victims were clearly aware of the facts, which had been widely publicised by the media with the launch of Operation Fanta on 24 January 2006, and could take the legal action they considered appropriate.

Therefore, the three-year limitation period in this case began on 24/02/2006, with the publication of the initiation of administrative proceedings at CADE, and expired on 24/02/2009, before the distribution of this action (05/03/2021).” [Emphasis supplied]

A similar conclusion was reached in Manzoni and Fabbri with the court in the latter case concluding that:

“… the decision that gave wide publicity to Operation Fanta and the existence of the proceedings at CADE date back to 2006 and constitute a notorious fact, with the actual possibility of knowledge of the imputation of practices by the defendant in the outbreak of the operation, and it is unreasonable to consider the long process of the proceedings at CADE as relevant.”

In consequence time started to run from then except in relation to contracts entered into after the date, when time started to run from the date when the contract was entered into.

The conclusions reached in each case rested on knowledge that it was concluded was known to the claimants or could reasonably have been obtained by them. As the court in Manzoni held, the facts and matters referred to above (the publicity attaching to Operation Fanta and the publication of the notice concerning the initiation of the administrative proceedings) were to be viewed together with the claimant’s presumed knowledge that “… upon contracting with the price adjustment, the claimant orange grower was aware of the market price and the cost of production and that he had not received a better offer, which is why he ended up doing business with the defendant.”

There was a dispute between the Brazilian law experts as what knowledge was required for present purposes. While both were agreed that what was required was unequivocal knowledge, they disagreed as to what a claimant was required to have unequivocal knowledge of before time started to run for limitation purposes.

Professor Frazão considered that unequivocal knowledge was required of at least the violation of the right (in this case the existence of the alleged cartel), authorship and extent of damages. Professor Didier considered a claimant was required only to have unequivocal knowledge of the possible violation of the claimants’ rights and possible authorship. He maintained that a claimant was not required to know the full extent of its possible damage. In my judgement on this issue Professor Didier’s evidence is to be preferred. I reach these conclusions for two reasons.

Firstly, as Professor Didier said in the course of his evidence, a claimant will necessarily not know with certainty that its rights have been violated or the extent of the loss that has been caused until after a trial. This is particularly so in relation to loss, which will involve a technically complex investigation as to what price movements would have been dictated by ordinary market conditions and the elimination of those movements from the calculation of loss caused by the alleged cartel activity. This is very often a highly complex exercise carried out by expert economists usually applying multiple regression analysis. An example of this is provided by Granville Technology Group Ltd v Chunghwa Picture Tubes Limited and others [2024] EWHC 13 (Comm). Although this was a claim governed by English law, the technical enquiry necessary to identify the loss attributable to alleged cartel activity will usually be broadly similar whichever law governs the claim. If it was to be concluded that time could not start to run until the claimant knew with certainty the loss that it had been caused the reality is that in most of not all cases time for limitation purposes would never start to run. Similarly, it is impossible for a claimant to be sure its rights have been violated. As was put to Professor Didier, even if the commencement of an investigation into an alleged cartel is a notorious fact, that does not mean that the existence of the alleged cartel is a notorious fact. I agree. That points firmly against a victim being required to certainly know its rights have been violated but only that they may have been. If that was not so, then again time would never start to run for limitation purposes since it is only after a trial and judgment that anything approaching certainty could result. Indeed, in most cases even then certainty will not be achieved because conclusions will have been reached applying the applicable onus and standard of proof tests.

Secondly, Professor Didier’s analysis more closely reflects the way in which the test is described in the judicial decisions so far considered. I have referred to some of the formulations above. By way of example, the Court of Appeal in Ramos held that “… the Appellant could not file a claim before it was aware of the possibility of a cartel involving the companies with which it had contracts.” [Emphasis supplied] and later formulated the date at which time started to run as being “… the point at which the potentially injured party became clearly aware of the alleged existence of the cartel and the possibility of having suffered damage and could then take legal action to preserve or enforce its rights…” Applying this test, the Court of Appeal concluded that:

“… the three-year limitation period for non- contractual civil damages (Art. 206, para 3, V, CC) should be adopted as the starting point, the date on which the opening of administrative proceedings at CADE against the appellee companies was published on 24 February 2006, the date from which the potential victims were clearly aware of the facts, which had been widely publicised by the media with the launch of Operation Fanta on 24 January 2006, and could take the legal action they considered appropriate. Therefore, the three-year limitation period in this case began on 24/02/2006, with the publication of the initiation of administrative proceedings at CADE, and expired on 24/02/2009, before the distribution of this action (05/03/2021).”

The appeal to the STJ failed.

A similar approach was adopted by the Court of Appeal in Manzoni where the Court of Appeal concluded that time started to run from “… the moment the claimant was unequivocally aware of the alleged existence of a cartel and the possibility of having suffered damage, his interest in claiming his right begins…” and that even though “… there is nothing in the file to show that he was aware of the existence of the cartel to support his claim for damages…” the court nonetheless concluded that “… the claimant unequivocally became aware of the formalisation of the discussion on the existence of a cartel in the citrus sector and, consequently, of the losses he had suffered…” because of “… the widespread media coverage of "Operation Fanta" in 2006, which investigated Brazilian companies in the orange juice sector and the formation of a cartel, as well as the initiation of administrative proceedings before CADE, published in the Official Gazette that same year (24/02/2006)…”.

These formulations are significant for at least two reasons. Firstly, they are inconsistent with Professor Frazão’s evidence concerning the requirement for certainty and secondly, they demonstrate unequivocally that Brazilian law approaches the issue of knowledge by reference to information that objectively was available even if it was subjectively unknown to a particular claimant.

For those reasons I conclude that what is required for time to start running is unequivocal knowledge on the part of the claimant of the possibility of a violation of its rights and the possible party or parties responsible. I reject the notion that unequivocal knowledge that the claimants’ rights have been certainly violated is required.

There is a dispute between the experts as to the degree of knowledge required of the loss caused by the alleged cartel. I reject the proposition that for time to start running the claimant must have unequivocal knowledge of the extent of the losses caused by the violation of its rights. It is implausible that no knowledge of loss is required for time to start running, if for no other reason than that a party would not commence legal action other than to recover a loss, but either that is part of the knowledge necessary to demonstrate a possible violation of rights or (if it is a separate requirement) it requires nothing more than unequivocal knowledge that some loss has or may have resulted from the alleged violation of rights. There are some cases decided by the STJ in which the knowledge required was described as being unequivocal knowledge of the extent of the damage caused. In my judgement that cannot sensibly be treated as meaning that time does not start to run until a claimant has knowledge of the whole of the loss that it has suffered. I have referred already to the complexities that apply to assessing loss in alleged cartel claims and if it was concluded that a claimant had to know with certainty the extent of the losses caused by being exposed to alleged cartel activity, then in many and perhaps most cases, the limitation period would never start to run. I reject therefore the claimants’ submission that for time to start running for limitation purposes what is required is knowledge of the full extent of the loss suffered by the claimant concerned.

In any event it is absurd to suppose this could be affected by a CADE final decision because that is concerned with an administrative investigation and the imposition of fines whether arrived at by agreement or otherwise. It is not at all concerned with private law claims against alleged cartelists or the losses suffered by particular victims as a result of alleged cartel activity. That is why even in a follow-on claim, a claimant has to prove causation and loss. As with the other knowledge requirements this can be established either by knowledge actually possessed by the claimant or can be inferred.

This material leads me to conclude that it is highly likely that having rejected the suggestion that time did not start to run until after publication of the CADE Final Decision because that decision is not condemnatory a Brazilian court would conclude that the claims being pursued in these proceedings were stand-alone claims that had been commenced long after the expiry of the relevant limitation period because such a court would conclude that time started to run by no later than February 2006, when (following Operation Fanta) notice was published of the widening of the regulator’s administrative proceedings. It would do so because, by that stage at the latest, the claimant would have known both of the possible existence of the alleged cartel and that its membership included SCL, being the processor to which the claimant concerned was forced to sell its product, and the defendants. It would have also known (assuming such to be the case) that the purchase price of oranges paid to the claimants was reducing with no alternative market available and/or that production and logistic costs borne by the claimants being increased. All this was knowledge clearly known to the claimant concerned because it was within his own knowledge by reason of the business he operated (as was the case in relation to prices and costs) or was notorious (as was the case with Operation Fanta and those who were subjected to the dawn raids, as to which see the newspaper reported referred to earlier) or knowledge that could reasonably have been discovered by such a claimant (as for example the notice concerning the administrative procedure by CADE as publicised by its notice published in the Official Gazette). Professor Frazão accepted that it was open to the courts to find that, after Operation Fanta and publication of the notification in February 2006 the claimants were (or ought reasonably to have been) aware of the alleged existence of the alleged cartel – see T3/38/2-8. She also accepted that at that point a claimant would be aware of the possibility of damage having been suffered - see T3/38/9-12. Her point was that more than mere suspicion was required. I have rejected that proposition for the reasons set out earlier.

In reality the information available by February 2006 was substantial. This material is summarised comprehensively of the List of Agreed Facts, and I have set out the details concerning most of the major events and sources earlier in this judgment. In summary however, Operation Fanta was a dawn raid against SCL amongst others and the fact of the raids and those targeted was widely reported – see the first of the newspaper extracts quoted earlier. It was also widely reported in Associtrus’ newsletters. This was followed on 24 February 2006 by the publication in the Official Gazette of the SDE’s notification that its enquiries had been expanded to include the first and second defendants amongst others as well as recording the names of those already being investigated. I have set out in detail earlier the content of Mr Machado’s testimony, Mr Borella’s testimony, the content of the criminal complaint and the Opinion of the Federal Prosecutors Office concerning the proposed 2006 TCC. The key point for present purposes is that each contained a significant amount of detail concerning the existence and operation of the alleged cartel. In my judgement the reasonable availability of this material is manifest from what was served with or pleaded in the Costa and Favero Proceedings referred to earlier. The former was brought on behalf of 46 claimants and the latter on behalf of 51 claimants. The material filed with each claim included the criminal complaint, Mr Machado’s testimony Mr Borella’s testimony and the Federal Prosecutor’s Opinion. As set out in more detail earlier, the criminal complaint relied on reproduced either verbatim or in summary the evidence of Mr Machado and Mr Borella, the Prosecutor’s Opinion sets out a summary of the evidence said to establish the continuation of the alleged cartel after 2000 including the reproduction in detailed summary of the evidence of some former employees of the alleged cartel. The claimants submit that this material does not assist because the Favero and Costa Proceedings (and the Machado Leniency Agreement) were concerned with events up to 2000 whereas the claims made in these proceedings run to 2006. I reject that as immaterial to the issues that arise, which concern available of reasonably acquirable knowledge as to the possible existence of the alleged cartel. These claims demonstrate the availability of such information.

This litigation was supported by Associtrus or at any rate the commencement of the proceedings was widely referred to in its newsletters. Examples include:

“Citricultors go to court to terminate contracts

30 November| 2006

Producers linked to the Brazilian Association of Citrus Growers (Associtrus) are going to court to have orange supply contracts cancelled. The producers must base the class actions on the decision of the Administrative Council for Economic Defence (Cade), taken last week, to stop the agreement that would have closed the investigation into cartel practices by the processors.

"This showed that the contracts were made in an environment in which there is evidence of cartel practice, which is why they can be cancelled," affirms Flávio Viegas, president of Associtrus. Today, the organisation is holding a meeting in the city of Bebedour to decide the course of these actions.

According to Viegas, the contracts that will have to be cancelled are those made up to These are five years out of date, with payments of between US$3 and US$3.50 per box (40.8kg) and have not been renewed. "There was no competition for the fruit," said the president of Associtrus.

Flavio Viegas, affirms that some citrus growers have already gone to court and one of them reached a one-off agreement with the industry, with payment of up to $5.50 a box…”.

In an article in its newsletter dated 10 November 2006, under a headline that referred to the dispute in 2006 as to whether a TCC should be entered into, under the strap line “Companies are accused of forming a cartel”, Associtrus identified by name the entities (including SCL) that would have been parties to the disputed 2006 TCC. It referred to the 1995 TCC, alleged that it had not been complied with and that it was this that caused “… a new investigation, which identified the continuity of the practice…” which resulted in the events referred to above as Operation Fanta.

By 5 February 2007, Associtrus was able to report in an article in its newsletter of that date a comment of the lawyer retained by Associtrus that “… lawsuits, with hundreds of producers, are being filed in the cities of São Paulo and Matao”. In relation to the information that was available, Associtrus was able to say in its newsletter of 26 March 2007, under the headline “Citrus farming – lies vs truth”, that “… it's been a long time since information has been available. necessary for the Citrus Grower are on the ASSOCITRUS website…” and that “…(t)oday, information is accessible to everyone and the mystery can be unravelled in just a few minutes”. Whilst there were numerous other articles published thereafter, it is necessary to refer to only one more: In its newsletter dated 2 May 2007, under the headline “Justice rejects Abecitrus' request to end of the cartel investigation” Associtrus reported a comment of Mr Viegas (then the President of Associtrus and the lead claimant in the lead action in these proceedings) that he wanted officials to analysis the available material more quickly because “… there are already enough documents to prove that a cartel has been formed”.

In light of this material, I conclude that time started to run on the basis of knowledge that could reasonably have been acquired no later than the date when Operation Fanta was reported in the press as described above (when in at least some of the reports those that had been subject to the dawn raids was publicised) and publication in the Diário Oficial of the notice of the expansion of the administrative proceedings. It named both defendants and SCL in these proceedings as persons in respect of whom there were “…indications in the file of conduct contrary to the economic order…”. It was at that point that knowledge of the possible existence of the alleged cartel and of those who were allegedly parties to it became publicly available information. The materials and events to which I have referred which post-date those events only serve to emphasise that was so.

This approach is one that has been consistently adopted by the STJ – see its judgments in the cases referred to above. In relation to loss, as I have decided already, for time to start running a claimant needs to know only of the possibility that loss has resulted from the harmful conduct that forms the basis of the cause of action available to that claimant. The possibility that loss will have been suffered necessarily follows from actual knowledge of the terms of the contract or contracts that the claimant has been forced to enter with the relevant processor. Where those contracts were entered into after the date when the information concerning the possible existence of the alleged cartel and the participants in it became publicly available, then time started to run from the date when those contracts were entered into because it was from that date such a claimant became aware of possible loss being caused by alleged cartel activity by that claimant’s processor counter party.

The CADE Final Decision takes matters no further because it is not condemnatory – that is it does not record any findings of anti-competitive conduct any more than the TCC contains any admission of cartel activity. These conclusions reflect the conclusions consistently reached by the STJ in its judgments in the numerous cases referred to above. The facts relevant to the acquisition of knowledge are as I have set out above. The effect of that material on limitation is a matter of Brazilian law and the issue being whether the STJ would arrive at a different conclusion by reference to this factual material if the issue was to come before it again. I can with confidence conclude that it would not. That is apparent not merely from the principal cases considered at the trial but the decision that followed.

For these reasons I conclude that time started to run for Brazilian limitation law purposes no later than 24 February 2006 or the date when the claimants entered into any relevant contract with a processor alleged to be part of the alleged cartel if later. This latter qualification is immaterial because in these proceedings as they are currently constituted, the cartel is alleged to have continued in operation until no earlier than 24 January 2006. The claim form in the Viegas Claim was issued on 27 September 2019 and the claim form in the Sanches Claim was issued on 22 November 2019. It follows that each of the claim forms was issued at a time when the claims were barred by operation of Brazilian limitation law.

Although the defendants suggested this approach was similar in effect to that adopted as a matter of English law when applying s.32 of the Limitation Act 1980 – see Gemalto Holding BV and others v Infineon Technologies AG and others [2022] EWCA Civ 782; [2022] 3 WLR 1141, in my judgement that is immaterial and I should make clear I have taken no account of that submission (or the decision of the Court of Appeal in Gemalto) in arriving at the conclusion I have arrived at here. Gemalto was concerned exclusively with English law limitation issues in a competition law context whereas this case is concerned exclusively with Brazilian law limitation issues in a competition law context.

This issue is or may be material to the application for permission to amend that the parties have agreed should not be determined at this stage. The issue concerns whether the claimants’ claims are to be characterised as claims in respect of continuous damage as a matter of Brazilian law and if so when the limitation period for delictual claims causing continuous damage starts to run as a matter of Brazilian law.

Professor Frazão maintains that where there is continuing misconduct resulting in permanent harm, time does not start to run until the cause of the harm has ceased. She further maintains that alleged cartel activity is continuing misconduct that results in personal harm with the result that time for limitation purposes does not start to run for any of the loss alleged to have been caused by continuing misconduct however long ago that loss might have been caused – see paragraphs 223-224 of her second report where she stated:

“223.

According to the STJ, the limitation period could only begin after cessation of the cause: “[t]herefore, there is no need to talk about a lapse of the limitation period in the case of constant and permanent damage that continues until the claim is filed. After all, if the damage stems from a cause that extends over time, it is from the cessation of the cause that the limitation period begins to run”.

224.

In the same sense, in an action for compensation for environmental damage, the STJ emphasised that “[t]here is no limitation, since the maintenance of buildings in the environmental preservation area prevents vegetation from regenerating, thus prolonging the If correct, then in relation to competition law claims time would not start to run in relation to losses suffered decades before. damage caused to the environment. In this case, the damage has been perpetuated, recreating or renewing the legal right to claim of the holder of the offended right every day. There is no lapse of the limitation period in environmental actions arising from permanent damage, at least as long as the environmental damage persists”.

In support of this proposition, Professor Frazão relied on the judgment of the STJ in Telemar Norte Leste SA v. Delgado Special Appeal No. 1.659.500 - RJ (2015/0007613-0). The issue in that case was whether a claim for compensation for “moral damage” suffered as a result of the alleged violation of neighbourhood rights by the appellant was time-barred. The underlying basis of the claim was that a plot of land next to the claimant’s home belonged to a utility provider but had not been secured with the result (so it was alleged) that it was “… used by passers-by, where they get dirty, use drugs, urinate and litter, making the place a tremendous open-air public restroom and trash dumping…” The appellant denied liability on the basis that it had not abandoned the land and that it routinely cleaned and otherwise maintained the plot and in any event the claim was time barred because “… the action was filed more than 3 (three) years after the alleged violation of the right.” It was held that no limitation issue arose because “… there is no need to talk about a lapse of the limitation period in the case of constant and permanent damage that continues until the claim is filed. After all, if the damage stems from a cause that extends over time, it is from the cessation of the cause that the limitation period begins to run.” The judge rapporteur continued:

“15.

… it is worth remembering that the situation would be different if the company or third parties had, on a single occasion, deposited rubbish on the land adjacent to the plaintiff's home, with the waste remaining in the area over the years. In this situation, the deadline for filing a reparation and compensation action should be counted from the date on which the neighbourhood right was violated, and there is no need to talk about renewing the limitation period on a daily basis.

16.

It should be noted that, although it does not portray a situation exactly similar to the one in this case, this panel of judges, when judging a dispute regarding the limitation period for the filing of an action for compensation by the beneficiary of housing insurance arising from construction defects in the property which, in turn, ended up causing various other damages to the property, such as infiltrations, cracks in the walls, etc., expressly recognised that the damage to the property was continuous and of a gradual nature, expressly recognised that the damage was continuous and permanent, concluding that "since the damage to the property is of a successive and gradual nature, its progression gives rise to numerous claims subject to insurance cover, renewing the insurance beneficiary's claim and, consequently, the starting point of the limitation period".

17.

As we have seen, although it does not reflect a situation identical to the present one, it clearly demonstrates the occurrence of a renewal of the limitation period when there is damage that extends over time, which is why the judgment under appeal does not deserve to be reformed on this point.”

Professor Didier said in the course of his evidence that he had not read this particular judgment which limited the evidence he could give about it. However, he did make one point in relation to the issue that I am now considering which was that if the law in relation to alleged cartel claims was as Professor Frazão asserted by reference to Telemar Norte Leste, then “we would never have limitation”. Since Professor Didier was not familiar with this case, he was not able to assist me further as to its scope and effect.

Professor Frazão was able to assist as to the scope and effect of the decision. Her point was that:

“… we have here is two different theses. If the damage is permanent, we need a cessation of the cause. However, if the damage is successive or periodic, we then can have the renewal of the limitation period every day, and this is an important case for our cause because we need to understand that the cartel is a permanent infraction, that is it involves neutralising the conditions for competition for some time, and that's why the conduct is permanent and the damage is permanent too…”.

Professor Frazão maintained that it was wrong to conclude that damage results from an alleged cartel only because of the sale contracts that the producer enters into with the processor. In her view:

“… market sharing is a clear example of permanent damage. That's why every producer can only sell their outputs to a specific purchaser and they are at the mercy of that buyer. This is a damage that's clearly permanent. So abusive contracts is only the consequence of a structure of a cartel which projects itself beyond these contracts. There are many damages by a cartel which go beyond the conclusion of contracts.”

This was in the end the difference between the experts on the issue I am now considering. Professor Didier was taken to the decision of the STJ in Manzoni referred to earlier. As is apparent from the text of the judgment of the court below reproduced in the STJ’s judgment, the claimant in that case maintained that he had been caused serious losses by reason of being required to sell his produce at uneconomic prices as a result of the alleged cartel and that he claimed as damages “… the difference between the amount actually paid for the orange box and that which should have been paid had the market operated without the cartel's interference.” In relation to limitation, the lower court had concluded that:

“Despite all disclosure, the claimant maintained a contract with the defendant after becoming aware of the cartel. Therefore, the statute of limitations on the subsequent contract began with the date of the contract.

However, even so, the initial claim had long been covered by the statute of limitations. This is because, even if the date of termination of the contract in 2010 is considered, the present claim was only filed in 2021, which is well beyond the three-year term.”

Professor Didier was taken in re-examination to the relevant paragraphs of Telemar Norte Leste and then to Manzoni and was asked “… to explain to the court, as briefly as you can, how limitation works where successive contracts are made with the Claimants pursuant to an alleged unlawful cartel…” to which he responded:

“When it's permanent, it generates several contracts, as is the case, the limitation period starts – it generates for each different contract. So each different contract generates damage. So prescription -- limitation starts on each contract to seek damages that each contract generated due to the formation of cartel, exactly as this extract is describing.”

I accept this part of Professor Didier’s evidence for the following reasons.

I do not accept the proposition that in a private law claim for damages by an alleged victim of an alleged downstream supply chain cartel, damages are likely to be assessed on anything other than the difference between the amount actually paid by the alleged downstream defendant cartelist to the upstream vendor or service provider claimant and the price that would probably have been paid had the relevant market operated free of the alleged cartel. One way of giving effect to a limitation period where there are continuing losses is to hold that all losses incurred prior to the date when time started to run are barred. A variant of that, where the losses are the result of long term supply contract is to consider losses flowing only from such contracts as are entered into within the relevant limitation period. In a case such as Manzoni therefore the losses suffered as a result of contracts entered into with a processor three years or less before the claim was commenced would in principle be recoverable providing all the relevant elements of the claim were established.

Professor Frazão accepted that this approach was the one that had been consistently adopted by the STJ in the alleged orange cartel cases considered above – see Teles, Jotto and Manzoni. There is no competition law case to which my attention was drawn where any attempt was made to rely on Telemar Norte Leste or the theory set out in it, much less one where the theory was adopted. That is significant in itself but is the more significant because the judge who decided Telemar Norte Leste was Andrighi J who was judge rapporteur in Fabbri and Neto and the panel that voted to adopt Andrighi J’s judgment in Telemar Norte Leste included Cueva J who had dissented in each of those cases. The approach adopted in the alleged orange cartel cases is consistent with loss due to the alleged cartel activity being successive or periodic loss as defined by Professor Frazão. Whilst what comes within the scope of “permanent” loss was not anywhere precisely defined by Professor Frazão, in my judgement the loss suffered by an upstream supply chain supplier selling to an alleged downstream cartelist is successive or periodic because a new loss is suffered every time a new supply contract is entered into.

For those reasons I reject the claimants’ case on the issue I am now considering.

For the reasons set out above, I conclude that the claims as presently pleaded are time barred and ought to be dismissed on that basis.