1. This study’s primary objective is to analyse how national courts in Europe have assessed cartel overcharges. In addition, it also provides figures on the development of cartel damages actions in Europe (how many cases were judged, in which countries, with which outcomes, etc.). It was completed with the help of lawyers, law professors, national competition authorities and national judges from 30 European countries. Judgments in cartel damages actions have been systematically identified, compiled, translated and analysed. This article is presenting results of this research.
2. Section I describes the methodology followed. Section II provides general figures on the cases gathered. Section III analyses the award of damages. Section IV focuses on how courts have assessed cartel overcharges, and also passing-on. Finally, section V presents highlights from some recent cases.
I. Research methodology
3. The research methodology for this year’s edition is largely similar to last year’s.
4. Scope. In this article, the term “cartel” has the meaning given by the European Commission: “A cartel is a group of similar, independent companies which join together to fix prices, to limit production or to share markets or customers between them.”  A “case” means an action for damages, with one or several plaintiffs alleging that a cartel caused an overcharge, and in which a court handed down at least one judgment on the merits. This includes three sets of judgments: judgments awarding damages, judgments establishing liability but not valuing the damages,  and judgments dismissing actions for lack of merit.
Some of the judgments analysed are not final. Sometimes this article refers also to judgments in cases other than cartel damages actions, or to judgments that are not judgments on the merits, when a specific part of their content is particularly interesting.
5. Importantly, cases in which an out-of-court settlement was reached before any judgment on the merits fall outside the scope of this study. Cases dismissed on strictly formal grounds such as jurisdiction or statute of limitations are not included either. 
6. Counting cases. Counting cases required setting a rule for this purpose. Sometimes several judgments are similar. For example, on 10 January 2020 the Provincial Court of Barcelona handed down three judgments on actions that followed the Spanish paper envelopes cartel.  These three judgments are counted as three distinct cases. When a large number of judgments are similar, however, an exception to this rule is made. For instance, on 20 October 2016, the Helsinki Court of Appeal gave 40 judgments in actions related to the Finnish asphalt cartel. Counting each of these judgments as an individual case would give them excessive weight relative to other cases. For this reason, each large set of relatively similar judgments is considered to represent a single case. 
This rule has an important effect in this edition of this study. By the end of 2020, Spanish courts had given hundreds of judgments in actions following the European Commission decision in case AT.39824 – Trucks; some say about 500 judgments.  In reading what follows, one needs to keep in mind that all these actions are counted as only one case. 
7. Geographic coverage, research period and temporal scope. This research covers the post-Brexit EU comprising 27 Member States, plus Norway, Switzerland and the United Kingdom. It was conducted for the most part between February and July 2021.
The temporal scope of this study runs from 30 June 1998, when the first judgment in our database was given, until 30 December 2020. Judgments handed down since 1 January 2021 are not included. 
8. Research process. The process employed for this research has four steps. The cases were identified. Copies of judgments were gathered. Using recent machine translation software, they were translated into English.  Their content was then analysed.
9. Contributors. This year again, contributors have played a critical role at all stages in this research. They are often lawyers or law professors. They were asked whether they were aware of relevant cases in their jurisdictions. Many helped identify such cases, and often assisted with their analysis. Many others indicated that there had not yet been a suitable case in their country.  Lists of cases were subsequently reviewed by national judges, and they were checked by national competition authorities (NCAs). 
In total, 113 lawyers, law professors and economists, 26 NCAs and 56 judges directly contributed to this study.
This research would not have been possible without the invaluable assistance of Fernando Aguilar de Carvalho, Philip Andrews, Tonia Antoniou, Anastasios A. Antoniou, Elena Apostolova, Sylvann Aquilina Zahra, Rasmus Asbjørnsen, Ján Augustín, Ieva Azanda, Georgiana Bădescu, Zoltan Barakonyi, Alessandro Bardanzellu, Daniel Barry, Jean-François Bellis, Ana Belén Blasco Montés, Miklós Boronkay, Mislav Bradvica, Helmut Brokelmann, Lauras Butkevičius, Davide Cacchioli, Rino Caiazzo, Antonio Campitiello, Maja Činč, Stamatis Drakakakis, Aleksandra Dziurkowska, Marc Felix, Thomas Funke, Nikolai Gouginski, Alessandro Greco, Eline Groen, Manuela Guia, Anna Gulińska, Franz Hoffet, Marek Holka, András Horváth, Sarah Houghton, Pavel Hristov, Smilena Hrusanova, Vilhelmiina Ihamäki, Isabelle Innerhofer, Toni Kalliokoski, Johan Karlsson, Matej Kavčič, Anikó Keller, Jiří Kindl, Thomas Knapowski, Mario Krka, Maria Lampadaki, Valérie Lefever, Raquel Sofia Lemos, Augustė Linauskaitė, Palle Bo Madsen, Martin Mäesalu, Richard Maliniak, Monika Mališauskaitė-Vaupšienė, Tomas Maretta, Carmen Martínez Mateo, Laura Mihalache, Gildas de Muizon, Annalies Muscat, Martin Nedelka, Robert Neruda, Irmantas Norkus, Jörg Nothdurft, Andreea Oprișan, Peter Oravec, Trine Osen Bergqvist, Raino Paron, Jasminka Pecotić Kaufman, Vladimir Penkov, Javier Pérez Fernández, Petra Joanna Pipková, Polina Polycarpou, Roman Prekop, Adam Přerovský, Mani Reinert, Risto Rüütel, Anders Ryssdal, Erik Söderlind, Miguel Sousa Ferro, Laura Spiteri, Aleksander Stawicki, Dragomir Stefanov, Agnieszka Stefanowicz-Barańska, Christian Steinle, Valeriu Stoica, Magnus Strand, Pedro Suárez, Daivis Švirinas, Elo Tamm, Stefan Thomas, Fabio Trevisan, Jon Turner QC, Stefan Tzakov, Dimitris Tzouganatos, Lumine van Uden, Raluca Vasilache, Weyer VerLoren van Themaat, Patricia Vidal Martínez, Otilia Vîlcu, Maaike Visser, Franziska Weber, Frank Wijckmans, Hanno Wollmann, Peter Wytinck, Petr Zákoucký, Janja Zaplotnik, Rasa Zaščiurinskaitė and Uģis Zeltiņš. 
I am very thankful to the competition authorities who have reviewed and enriched lists of cases, including the Austrian Federal Competition Authority, the Commission on Protection of Competition of Bulgaria, the Croatian Competition Agency, the Commission for the Protection of Competition of the Republic of Cyprus, the Office for the Protection of Competition of the Czech Republic, the Estonian Competition Authority, the Finnish Competition and Consumer Authority, the French Autorité de la Concurrence, the German Bundeskartellamt, the Hellenic Competition Commission, the Irish Competition and Consumer Protection Commission, the Competition Council of Latvia, the Lithuanian Competition Council, the Luxembourg Competition Council, the Office for Competition of Malta, the Norwegian Competition Authority, the Polish Office of Competition and Consumer Protection, the Antimonopoly Office of the Slovak Republic, and the Spanish National Commission on Markets and Competition.
I would like to express my deep gratitude to the judges who have kindly contributed to this research, including Joana Manuel Mateus Araújo, Elske Boerwinkel, Marc Bosmans, Mads Bundgaard Larsen, Marta Borges Campos, Paolo Catallozzi, Mark Chetcuti, Angelos David, Guido de Croock, Mieke Dudok Van Heel, Ulrich Egger, Jürgen Exner, Karin Fløistad, Silvia Giani, Katalin Gombos, Petra Hočevar, Thierry Hoscheit, Irmantas Jarukaitis, Wolfgang Kirchhoff, Gerhard Klumpe, Villem Lapimaa, Irène Luc, Purificación Martorell Zulueta, Liam McKechnie, Christina von Merveldt, Vanda Miguel, Krasimira Milachkova, Polona Mlakar Adam, Tibor Tamás Molnár, Andrea Moravčíková, Cristian Daniel Oana, Maria Arántzazu Ortiz González, Eduardo Pastor Martínez, Maria Mercedes Francisca Pedraz Calvo, Igor Periša, Rūta Petkuvienė, Sylvaine Poillot-Peruzzetto, Maja Praljak, Mira Raycheva, Jacqueline Riffault-Silk, Samuel Rybnikár, Tomáš Rychlý, Adam Scott, Ingeborg Simonsson, Ewa Stefańska, Iannis Symplis, Rudīte Vīduša, Sabine Voelkl-Torggler, Jaap de Wildt and Anne-Marie Witters. Other judges have also contributed to this study and are not named here.
10. European Commission and Association of European Competition Law Judges (AECLJ). Finally, I am grateful to the European Commission for having expressed interest in this study and for our fruitful exchanges. I am further indebted to the Association of European Competition Law Judges (AECLJ) for its encouragement and its non-financial support.
11. Limitations. This research is subject to three main limitations. First, the list of cases identified is, despite best efforts, unlikely to be completely exhaustive. Many cartel damages actions receive only very limited attention. On several occasions, contributors have uncovered judgments that were not publicized, cannot be accessed online, and have so far remained unnoticed. Given the wide scope of this research, however, some cases may not have been identified. I would be grateful to anyone who could bring to my attention any case of which I may not be aware.
12. Secondly, errors in interpreting the content of some judgments have possibly been made. The variety of languages in Europe constitutes, of course, a difficulty for this research. Most judgments gathered are neither in English nor in French.  In order to grasp some of their substance, a number of sources were used, including machine translations, expert analysis from contributors or articles describing the content of some of the judgments.  But as I could not read the original (untranslated) text of many judgments, I cannot exclude the possibility that I may have misunderstood part of their content.
13. Third, the figures provided in this study should be considered only as indicative. The judgments analysed represent in total well over 10,000 pages. Many criteria have been screened in each judgment—whether there was any mention of passing-on, of umbrella pricing, of a court-appointed expert, etc. At times, I have probably failed to notice the presence of some criteria in some judgments. 
14. Observations made in previous editions. Finally, a number of observations reported in previous editions of this study are still valid. They are often not repeated in this article.
II. General figures
15. Number of cases. In the 30 European countries covered, 299 cartel damages actions have been identified. They include 58 cases in which damages were awarded, 93 cases in which liability was established, and 134 cases that resulted in dismissals. Fourteen cases are also pending, often after a judgment was quashed. 
When interpreting these figures, it is important to have in mind that all damages actions, which were judged by Spanish courts following the European Commission decision AT.39824 – Trucks, are accounted for as only one case (see ¶ 6).
There were 45 cases judged for the first time in 2020, compared with 56 in 2019 and 53 in 2018. It seems that the number of recent cases was affected by the pandemic. In several countries, contributors have indicated that it has caused judgments to be delayed.
Figure 1. Cumulative number of cases, by date of first judgment
16. Number of judgments. The 299 cases represent in total 472 relevant judgments. As a benchmark, back in December 2009, when the study prepared for the European Commission  was published, the cumulative number of relevant judgments was 46.
The judgments were given by courts of first instance (314 judgments), courts of appeal (119) and supreme courts (39). Nearly half the supreme courts’ judgments were given in the last two years. On assessing cartel damages, in particular, several recent judgments handed down by the Bundesgerichtshof provide important guidance. 
As part of this study, 462 judgments have been collected, translated if necessary, and analysed. 
17. Countries. The cases come from fourteen countries: Germany (177 cases), France (52 cases),  Spain (25 cases), Hungary (8 cases), Italy and the Netherlands (6 cases each),  Belgium (5 cases), Austria, Finland and Greece (4 cases each), Denmark (3 cases), Poland and the United Kingdom (2 cases each), and Portugal (1 case). 
Once again, the number of cases would be much higher for Spain, absent the convention that all Trucks cases that were judged by Spanish courts are accounted for as a single unit.
18. United Kingdom. The small number of cases from the UK does not reveal the true share of London as a competition litigation forum. Many cartel damages actions have been brought before the Competition Appeal Tribunal or the High Court of Justice of England and Wales, in particular actions following European Commission decisions.  Most such actions were, however, settled before any judgment on the merits. “To our surprise,” wrote the Court of Appeal of England and Wales in a judgment handed down in October 2019, “we were told that this is the first case in which damages have been awarded by an English court or tribunal after a trial for anti-competitive cartel conduct which infringes Article 101(1) TFEU.”  Nevertheless, important insights on how to assess overcharges or passing-on can be found in various documents provided by British courts, including the judgments accounted for in this study, and also some judgments in cases in which the infringement was not a cartel.
19. Other countries. In a number of other countries, national courts have not yet judged on the merits any cartel damages action, but related developments have been noticed.
- In the Czech Republic, in May 2020, the Municipal Court in Prague discontinued proceedings against two defendants in a damages action that followed an infringement decision of the Office for the Protection of Competition (ÚOHS), after an out-of-court settlement was reached. In July 2020, the Regional Court in Prague approved an in-court settlement in another action.
- In Lithuania, in November 2019, a claim was submitted based on an infringement decision of the Lithuanian Competition Council. On 17 December 2020, the Supreme Court of Lithuania ruled on another case, which does not fit the definition of a cartel damages action given earlier, but in which reference was made to another infringement decision of the Lithuanian Competition Council. 
- In Norway, on 27 November 2019, the Supreme Court ruled on a matter of jurisdiction in a claim which follows an infringement decision of the European Commission. 
- In Romania, on 6 August 2020, the 6th civil division of the Bucharest Tribunal found a claim following a European Commission decision to be time-barred.  On 27 November 2020, the same tribunal rejected another claim as inadmissible. 
- In Slovakia, several proceedings were terminated by the District Court Bratislava II for procedural reasons, including issues related to the payment of court fees.
- Finally, in Sweden, following an infringement decision of the European Commission, actions for a negative declaration of liability (i.e. non-liability) were brought before the Patent and Market Court. 
20. Infringement decisions. Of the 299 cases, 57% followed an infringement decision made by a national competition authority, 40% followed a European Commission decision, and only 2% were stand-alone actions.  Most stand-alone cases correspond to civil actions brought before French criminal courts. In Italy, there is one case in which a national court referred to an infringement decision taken in another Member State.
National courts in Europe have decided on cartel damages actions that followed at least 72 infringement decisions (out of which 18 European Commission decisions).  They have thus developed insights on assessing cartel overcharges in a wide range of sectors and circumstances. However, it must also be recognized that this rich body of knowledge is geographically fragmented. In countries where at least one cartel damages action was judged, on average national courts have ruled in cases relating to (only) six distinct infringement decisions.  In other words, when dealing with future cases, judges and practitioners should often be able to uncover relevant past experience; but they will often find it in judgments given in a different Member State. Indeed, this study came across a significant and growing number of judgments, handed down in particular by Dutch, German or Spanish courts, making explicit reference to judgments given by foreign jurisdictions. 
21. Claimants: direct and indirect purchasers. About 72% of the actions were brought by direct purchasers.  However, the proportion of cases brought by indirect purchasers is now about 20%. This number includes a sizeable proportion of cases in which the claimants’ purchases were partly direct and partly indirect.
22. Sectors of claimants. Privately owned companies initiated 48% of the claims. Many other claimants come from a broadly defined public sector. Publicly owned companies (20% of the cases), local authorities (19%), and central governments (3%) initiated in total 42% of the claims.  Finally, 15 claims were brought by end consumers, 6 by claims vehicles, and 10 by various other types of claimants.
III. Damages awards
23. Damages awards. Cartel damages have been awarded in 58 cases.  They come from France (22 cases), Spain (16 cases), Germany (10 cases), Denmark (3 cases), Greece (2 cases), Austria, Belgium, Finland, Italy and the UK (1 case each).
Particularly in Germany, the number of damages awards is slightly misleading. In about 70 other cases, German courts have handed down interlocutory or declaratory judgments in which they have affirmed the liability of defendants without quantifying damages. 
24. Rate of success. The rate of success of judgments has evolved significantly in recent years. For this analysis, a judgment is considered successful if damages were awarded or liability was established—in other words, from the perspective of the claimants.  As shown in figure 2, from 2013 to 2018 the proportion of successful judgments has increased. In 2019 and 2020, it has dropped.
Figure 2. Outcomes of judgments, by year of judgment
Changes in the rate of success appear to be determined, at least partially, by the nature of the cases judged. Between 2015 and 2018, there were two large waves of judgments, one following the road signs cartel in France and the other following the rail cartel in Germany. Judgments belonging to these two waves have often been successful.
In interpreting these figures, again one needs to have in mind that they include for Spain only one case following the European Commission decision on Trucks. 
25. Figures on overcharges.  A rate of overcharge could be calculated or estimated for each of the 58 awards of damages.  In the past, this study used to indicate the individual rate of overcharge found in each case. This year, it is indicating the average rate of overcharge found for each cartel.  The range of overcharges is shown in figure 3. In accordance with the study prepared for the European Commission (Oxera et al., 2009), they are presented as a percentage of affected prices.  The lowest overcharge is less than 1%, and the highest reaches 34%.
Figure 3. Cartel overcharges in damages awards
The average rate of overcharge per cartel is 12%, and the median is 10%. 
IV. Assessments of overcharges and passing-on
26. Practical Guide. The European Commission published in June 2013 the Practical Guide on Quantifying Harm in Actions for Damages. This document describes methods considered by the Commission to be potentially suitable for assessing damages caused by competition law infringements.  It is referred to in many judgments. In some instances, judges have rejected quantifications of damages for the reason that these quantifications employed methods that were not listed in the Practical Guide. 
27. Methods accepted by courts. Courts have been exposed to all major types of methods described in the Practical Guide. In the 58 damages awards, damages were quantified with the following methods: 
- Comparison over time (also called “before-and-after”): 22 cases 
- Comparison with an unaffected market (also called “yardstick”): 6 cases
- Cost-based and financial methods: 9 cases
- Regression analysis (also called “econometrics”): 4 cases
- Simulation model: 0 cases
- Other methods: 23 cases 
28. Comparison over time. The method most frequently accepted by courts consists of comparing prices over time. This approach was noticed in particular in Austrian, French, German, Italian and Spanish judgments.
When a comparison over time is employed, where is the counterfactual price to be found: before or after the infringement? It was found after the infringement in about 70% of the cases. It was obtained before in some specific situations, for example when the cartel had a short duration. In some cases, before and after prices were used jointly.
In a judgment given in April 2020, the Commercial Court of Vienna observed that comparing prices over time could be difficult in, inter alia, two sets of circumstances: when order quantities varied and when technical innovations happened.
29. Comparisons with unaffected markets. Comparisons with unaffected markets are not frequently employed in cartel damages actions. In a small number of cases, claimants found a counterfactual scenario in a different country. But courts often challenged whether the choice of that particular country was truly appropriate. Recently, for example, the Commercial Court of Pontevedra questioned the suitability of a counterfactual found in Mexico, considering that the US or Japanese markets appeared to have more in common with those in Europe. Similarly, the Athens Court of Appeal considered that a comparison of prices in Greece and Germany was not conclusive, observing for example that input costs could differ. In practice, most comparisons with unaffected markets that have been accepted have stayed within national boundaries, comparing for example prices from different regions of the same country, or sometimes changes in prices for different product categories.
30. Regression analysis.  Past editions of this study could not find any damages award in which damages were quantified with regression analysis. There is some change. In February 2020, the Provincial Court of Madrid accepted one of the results of a regression model submitted by a defendant.  Although they are not accounted for in this study’s figures, there are also judgments in Spanish Trucks cases in which courts accepted the results of regression models.  In a set of recent judgments, a Spanish court has analysed distinct regression models providing diverging results. 
31. Simulation model. Judgments mentioning simulation models are very few.  This year, however, one such reference was found in a set of cases judged by the Regional Court of Cologne.  The court sought an expert opinion and, on the basis of extensive economic analysis, dismissed the claims.
32. Other methods. In 14 cases, courts have estimated a rate or an amount of overcharge. These cases come from Belgium, Germany, Greece and Spain.  Several courts have explained that they considered this method appropriate in situations in which the cost of obtaining the opinion of a court-appointed expert would be disproportionate relative to the potential amount of damages. It seems that this method was, in particular, frequently utilized by Spanish courts having valued damages in Trucks cases. 
In four German cases, courts have accepted the use of a predefined percentage stipulated in a contractual clause.  In some other cases, courts have referred to figures found in witness statements, to a number retrieved from an infringement decision, to the profit made by the infringer, or to the amount paid by the winner of a rigged bid to another participant in the tender.
Hungary and Latvia have established a rebuttable presumption that cartels cause an overcharge of 10% (or 20% in Romania).  This research did not come across any damages award based on such a presumption. However, in several Hungarian cases, the amounts of damages claimed were determined by reference to the presumption .
33. Requirements for reliable results. In a judgment handed down in December 2020, the Provincial Court of Valencia stressed that “it is not sufficient to use an accepted method for quantifying damages. It is also necessary that the data employed is sound, comparable and correctly processed. If such is not the case, the results obtained are either incorrect or biased.” 
34. Variations in the rate of overcharge. In December 2018, the Bundesgerichtshof observed that “cartel agreements which have existed for a long period and are meant to cover a large geographic area are likely to be of varying intensity over time and space.”  Some recent judgments are taking into consideration variations in the rate of overcharge over time.  Readers interested in this topic will also find data in two slightly older judgments, one handed down by the Paris Court of Appeal and the other by the Maritime and Commercial Court in Copenhagen. 
35. Proof of purchases. In cartel damages actions, evaluating damages often requires multiplying an amount of affected purchases by a percentage of overcharge. Most of the literature on quantifying harm deals with the second parameter—namely, the rate of overcharge. However, in a number of recent cases, defining and proving the amount of affected purchases were also central issues. In January 2020, the Regional Court of Stuttgart indicated for example that a claimant was expected to indicate “who procured from whom, when, which concrete object, in which manner (e.g., purchase, leasing, hire purchase) and at which conditions, in particular at which price.”  In July 2020, the Regional Court of Dortmund also indicated that allegedly affected transactions needed to be substantiated with regard to “the seller, the price and the exact product purchased.” 
36. Expert opinions. Several recent judgments provide considerations on how the courts can best utilize expert opinions. The Bundesgerichtshof reminded that no expert opinion, whether from a party-appointed or court-appointed expert, can replace the judge’s assessment of the case.  The Provincial Court of Barcelona observed that, when confronted with several conflicting expert opinions, the role of the judge is not to choose one against the others, but to shape his or her own assessment.  The Provincial Court of Valencia listed criteria to weigh the probative value of an expert opinion, including “the importance of the expert’s qualifications, his or her knowledge of the market, the method chosen and the reasoning behind his or her conclusions.”  The Paris Court of Appeal published sets of guidelines to be followed by court-appointed and party-appointed experts. 
In an important judgment handed down in November 2013, the Spanish Supreme Court indicated that any expert report shall “start from the correct bases.”  On this ground, Spanish courts have criticized expert reports for failing to stick to the findings of the relevant infringement decision.  A related objection was sometimes formulated by German courts. 
37. Umbrella effects. Two judgments provide criteria on how to assess umbrella effects—whether such effects appear likely or not, and whether their level should be expected to be high or low. One of these judgments was given in May 2019 by the Higher Regional Court of Düsseldorf, and the other in May 2020 by the Bundesgerichtshof. 
In an uncommon fashion, it was pleaded in a set of German cases that an umbrella effect had pushed up the prices not only of the cartelized goods, but also of substitutable products.
38. Passing-on. Passing-on was raised as a defence in approximately half the cases. This is a high proportion, having in mind that some cases are by nature ill-suited for invoking passing-on (for example when the claimant is a local government or an end consumer).
39. Passing-on in supreme court cases. In Germany, in a series of important judgments, the Bundesgerichtshof gave guidance on when and how passing-on should be taken into account when assessing damages.  One of these judgments is also dealing with the still relatively unexplored subject of mitigation through subsidies. 
In the UK, the Supreme Court stressed that “[t]he relevant question is a factual question: has the claimant in the course of its business recovered from others the costs of the MSC, including the overcharge contained therein? (. . .) If the court were to conclude on the evidence that the merchant had by reducing the cost of its supplies or by the pass-on of the cost to its customers (. . .) transferred all or part of its loss to others, its true loss would not be the prima facie measure of the overcharge but a lesser sum.”  The judgment confirms that in the event a merchant would react to an overcharge by cutting discretionary expenditure, reducing for example its marketing budget or its capital spending, the court would not consider damages to be mitigated. It also indicates that sometimes evidence on mitigation and passing-on is to be found in analysing the claimant’s budgeting process. Reference is made to this judgment in two recent judgments of the Competition Appeal Tribunal. 
40. Volume effect.  In May 2018, in Hungary the Metropolitan Tribunal dismissed a claim, considering inter alia that any overcharge had probably been passed on. In December 2019, the Metropolitan Court of Appeal ordered to re-examine this point: “If the injured party has passed on the overcharge partially or completely to its own purchasers, and has thereby reduced the incurred losses, the passed-on loss does not constitute anymore for this party damages for which compensation can be awarded. However, if it has passed on such damages, the price increase applied by the direct buyer probably leads to a decrease in the volume sold, which could constitute a loss of profit.” 
41. Passing-on to taxpayers. When a claim was brought by a central or a local government, occasionally courts have assessed whether such plaintiffs had possibly mitigated any overcharge through tax revenues. The Regional Court of Stuttgart has for example recently dealt with this question. 
42. Duration of cases and interest. In the 299 cases, the infringement decision came on average 8.4 years after the date of purchases; and the first civil judgment was handed down 4.5 years later. The total duration from the time the harm potentially occurred to the first judgment is therefore 12.9 years on average. Variations range, however, from less than 3 years to more than 20. This long duration explains why prejudgment interest is in many cases an important topic. 
V. Highlights from recent cases
43. In total, 127 judgments handed down between June 2019 and December 2020 have been compiled and analysed . Of course, this section can only provide a very limited overview of such rich material.
Appeals were possibly lodged against some of the decisions mentioned below.
44. On 18 June 2019, the Supreme Court of Finland dealt in two cases with the question of allocating the responsibility to pay damages to cartel members.  The court ruled that such allocation should be based on a combination of two criteria: which share of the advantage generated by the infringement had directly benefited the liable party, and which degree of fault was attributable to each party liable to pay compensation.  Applying these criteria, the court held one cartel member liable for one third of the damage in one case, and for the entire damage in another.
45. On 11 September 2019, the Amsterdam District Court asked litigation vehicles to further substantiate their claims.  As recalled in this interim judgment, there is now a number of Dutch court decisions addressing this particular subject. In this case, the court indicated that “the calculation of damages—including the underlying principles that should be taken into account—cannot be regarded as a ‘sub-topic’ in the phased handling of these proceedings, about which a separate ‘subordinate’ debate can take place. The calculation of damages concerns the core of the case.” 
46. On 23 October 2019, the Rotterdam District Court handed down another interim judgment dealing with a related matter. “The Court considers that X..., as assignee, has submitted a bundle of separate claims to the Court. This manner of proceeding does not allow for lower requirements to be set on the burden of proof than if each individual claim holder would have brought his or her claim separately. As a result, the possibility of damages of each individual claim holder needs to be established.” 
47. On 31 October 2019, the Court of Appeal of England and Wales gave its first judgment in a cartel damages action.  Readers will find in the sequence of judgments handed down in this case insights on a number of topics.  Three subjects covered by the Court of Appeal’s judgment deserve particular interest. First, the court agreed that “the judge was right to start without any presumption of loss or damage (. . .). We also agree with him that, on the facts of the present case, it is hard to see how such a presumption could have assisted [the claimant], given the need for its loss to be quantified and the generous approach adopted by English law to difficulties of proof in such a context.”  In particular, “the effect of an allocation-based cartel need not always be that a tender is made at an uncompetitive price.”  Second, the court explained that benefits obtained by a supplier from participating in a cartel are one thing; damages suffered by a client of this supplier can be another. “[T]he award of damages on the basis of savings made by the cartelist, rather than loss to the victim of the cartel as a result of having paid a price which was inflated by the conduct of the cartel, is based upon an error of law.” Third, as this study has observed in the past, assessing potential damages caused by bid-rigging raises a host of specific issues, several of which untouched in existing guidelines.  This judgment is providing reflections on some of these issues.
48. On 30 December 2019, the Commercial Court of Valencia No. 3 awarded damages in the amount claimed by the plaintiff.  In the case, the expert report submitted by the plaintiff included two distinct quantifications of damages, both of which seemed to provide relatively consistent results. Across all cases covered by this research, the ones in which one claimant presented several parallel approaches to quantify alleged damages appear to be relatively infrequent. However, it is perhaps interesting to notice that courts have sometimes tasked themselves an expert with, for example, “describing the different methods that could be employed for valuing damages, and employing at least two so that estimates are reinforced.” 
49. On 3 February 2020, the Provincial Court of Madrid awarded damages in two cases that followed the Spanish paper envelopes cartel.  Damages in these cases were quantified with regression analysis. Very uncommonly, the rate of overcharge that the court accepted was found in an expert report submitted by one of the defendants.
50. On 20 February 2020, the Athens Court of Appeal upheld a judgment of the Athens Court of First Instance awarding damages to three producers of milk.  This judgment deserves attention for at least two reasons. First, there are few cases in Europe in which sellers obtained damages from members of a cartel of buyers. Second, confirming the Court of First Instance’s holding, the Court of Appeal considered that claimants had incurred non-pecuniary (moral) damages. Moral damages are rarely awarded in cartel damages actions; outside Greece there appear to be only two other instances. Some readers might also be interested in the list of criteria that the court took into account in order to determine what would have been the “fair” purchase price. 
51. On 27 April 2020, the French Council of State  handed down a judgment on the financial consequences of nullity. “When a contract is nullified due to an anticompetitive practice attributable to the other contracting party, this party must refund all amounts paid by the public entity (having initiated the action), but it can ask in return (. . .) to be reimbursed for expenses that this party has made and that were useful to the public entity, excluding as a result any profit margin.” 
52. On 30 September 2020, the Regional Court of Dortmund awarded damages based on an estimated rate of overcharge of 15%.  In order to reach this estimate, the court observed inter alia a number of criteria that it considered indicative of the cartel effectiveness, including its duration, the cumulative market share of its members, the perceived level of cartel discipline, and whether or not there were substitutes to cartelized goods. Once such observations made, the rate of 15% was chosen by reference to several sets of information, including a predefined percentage found in general terms and conditions for one of the transactions, economic studies providing statistics on cartel overcharges,  and overcharges found by other national courts in Europe in other actions. The judgment also includes potentially useful developments on umbrella effects, and on requirements to be fulfilled according to the German Civil Code of Procedure by expert reports submitted by one party.
53. On 4 November 2020, the Regional Court of Dortmund estimated rates of overcharges in a second case. In this instance, the court valued the rates at 10% in the early years of the infringement and 15% later. 
54. To conclude, I would like to thank once again every person and every institution who offered their time and insights to this research.