In a bundle of 15 cassation judgments,1ECLI:ES:TS:2023:2472; ECLI:ES:TS:2023:2473;ECLI:ES:TS:2023:2474;ECLI:ES:TS:2023:2475; ECLI:ES:TS:2023:2476;ECLI:ES:TS:2023:2477;ECLI:ES:TS:2023:2478;ECLI:ES:TS:2023:2479;ECLI:ES:TS:2023:2480;ECLI:ES:TS:2023:2492;ECLI:ES:TS:2023:2493;ECLI:ES:TS:2023:2494. the Tribunal Supremo, the highest civil court in Spain (hereafter “TS”), rules on important questions concerning the ability of the judge to estimate the damage, the need for prior inter partes disclosure, proportionality, and standards for economic expert opinions. The TS does so against the background of the European trucks cartel, the same case that prompted the preliminary ruling in Tráficos Manuel Ferrer.2ECLI:EU:C:2023:99. The TS largely upholds the findings of the appellate courts and strengthens the ability of judges to estimate damage. The court also rules, among other questions, on the accrual of interest and limitation.
The TS judgments are issued against the background of a flood of litigation seeking damages against the members of the trucks cartel across Spain. Professor Francisco Marcos (IE Law School) is tirelessly compiling statistics on the resulting judgments. According to his latest LinkedIn post, made on the 26th of June 2023, there are by now 2,259 second instance judgments.3For a citeable source, see Francisco Marcos, Trucks cartel damage claims: thousands and odd judgements issued by Spanish Appeal Courts, Zeitschrift für Europäisches Privatrecht 1/2023. However, for the latest numbers, his LinkedIn posts are the better source. On 20th of June 2023, he posted a pie chart breaking down the results of these judgments to date; we re-create it below to provide a rough overview of the situation:
After some standardisation of results in the second instance,4The first instance judgements still pronounced overcharge percentages between zero (damage not proven) and 15%. the situation in these first 15 cases that reached the TS was as follows: The defendants’ economic expert reports had been dismissed as unconvincing. In 14 of the cases, the courts also dismissed the expert reports submitted by the claimants as insufficient. In these 14 cases, the lower instance judges then assessed characteristics of the cartel. While the product heterogeneity and individual negotiations might make collusion more complex, the long duration (14 years), the high market share of the cartelists (over 90%) and the wide geographic scope (the entire EEA) all militate for a strong effect of the infringement. Based on these and other findings contained in the decision of the European Commission, the judges estimated a cartel overcharge of 5% as a conservative “minimum damage”. A 5% overcharge estimation was also the outcome in about 47% of all second instance judgments (lightest green slice in the pie chart).
In the case ECLI:ES:TS:2023:2475, the appellate (second instance) court had instead upheld the first instance court’s judgment awarding damages based on an overcharge of 15% in line with the claimant’s expert report. This case hence belongs to the (heterogenous) slice with the darkest green colour.
a) Defendants’ expert reports deemed unconvincing
In these 15 decisions, the TS upholds the lower courts’ rejections of the defendants’ reports as a basis for quantification. In particular, the TS decidedly repudiates attempts by the defendants to re-classify the cartel as a mere information exchange and/or to argue that an effect on final transaction prices was impossible due to the distribution mechanics, discounts, or individual negotiations. That at least some reports by the defendants’ experts assume a mere information exchange, and hence start from an incorrect premise, is one of the reasons for disregarding them.
As many Spanish second instance courts, the TS cites the trucks cartel judgement of the Amsterdam court in CDC et al v DAF Trucks et al,5ECLI:NL:RBAMS:2021:2391, Rechtbank Amsterdam of 20 May 2021 RETAIL CARTEL DAMAGE CLAIMS S.A et al. vs. DAF Trucks N.V. et al. specifically, the image of the cartel as a rising tide that lifts all boats (prices), which was employed by CDC’s experts Professors Schinkel and Harrington.6Professor Maarten Pieter Schinkel, University of Amsterdam; Professor Joseph Harrington, University of Pennsylvania. Customer and transaction specific factors, such as buyer power, would also have been present without the cartel. These individual variations are like waves that exist in low tide and high tide but do not change the fact that, on average, a boat is higher in high tide than in low tide. The TS rejects the argument of the defendants that there was no causal link between the so called “list prices”7The term “list prices“ sounds as if these were public. However, that was not the case: They were strictly internal to each manufacturer, except for the illegal collusive exchanges and agreements. and the final transaction prices. According to the TS, it is implausible that top executives would, over a period of 14 years, spend time in several meetings per year discussing and agreeing on an irrelevant metric, all the while risking record-breaking fines for doing so. It is equally implausible that the lower distribution levels would have essentially absorbed the effect of the collusion during 14 years by decreasing their own remuneration.
Specifically, the TS makes more detailed observations about the defendant’s expert opinion submitted in ECLI:ES:TS:2023:2480. The mere fact that an expert report arrives at a very low overcharge percentage, and even a zero percent overcharge like in the case at hand, does not disqualify the report, as this would be tantamount to an irrebuttable presumption of overcharge. The same is true for the fact that it is based on a private dataset instead of public sources, although, if there were biased selection of data or similar, this would disqualify the opinion, an aspect however, about which the lower courts did not make any findings. But there are other problems with the data used for the expert report, namely that the data starts only in 2003 and ends in 2016. That the period from 1997-2002 is omitted is likely to introduce bias or lead to missing an important part of the effect. Even more interesting is the argument regarding the end of the dataset: The European Commission stated in paragraph 102 of its decision (dating from 2016) that it cannot be sure that the infringement has indeed completely stopped and ordered the companies to stop the infringement if they had not already done so. Hence, the period until 2016 cannot be viewed as a completely untainted, suitable control period. For all these reasons, the lower courts rejected the defendant’s expert report as unconvincing, and the TS confirms that the courts exercised their judicial discretion correctly in doing so.
b) Claimants’ reports deemed inadequate
In all of the decided cases, including ECLI:ES:TS:2023:2475, the claimant did not collect transaction data of a suitable quality and quantity to conduct case specific econometrics. Instead, the reports were “generic”. By reference to the datasets on cartel overcharges compiled by Connor, Lande, and Bolotova8Professor John M. Connor of Purdue University and the American Antitrust Institute has published on cartel overcharges for two decades. His SSRN page lists 99 scholarly papers, almost all concerned with cartels. The first publication listed that presents data on overcharges dates from November 2003 (Connor, John M., Private International Cartels: Effectiveness, Welfare, and Anticartel Enforcement (November 2003). Purdue Agricultural Economics Working Paper No. 03-12, Available at SSRN: https://ssrn.com/abstract=611909 or http://dx.doi.org/10.2139/ssrn.611909), the latest from January 2023 (Connor, John M. and Lande, Robert H., The Prevalence and Injuriousness of Cartels Worldwide ( 2023). Elgar Research Handbook on Cartels (Peter Whelan editor; Edward Elgar Publishing Ltd, United Kingdom) (2023 Forthcoming), University of Baltimore School of Law Legal Studies Research Paper Forthcoming, Available at SSRN: https://ssrn.com/abstract=4319572). Two frequent collaborators and co-authors of his are Professor Robert H. Lande (University of Baltimore) and Professor Yuliya Bolotova (University of Idaho), each of whom has also published empirical work on cartel overcharges on their own or with other co-authors. and/or literature and studies that analyse this dataset or certain subsets thereof (most famously the Oxera study9Oxera Consulting: Quantifying antitrust damages. Towards non-binding guidance for courts, Study prepared for the European Commission, ISBN 978-92-79-14685-5, doi 10.2763/36577 (pdf). and Smuda10Florian Smuda: Cartel Overcharges and the Deterrent Effect of EU Competition Law, ZEW Discussion Paper No. 12-050.), they argued for an overcharge percentage of 15% or 20%. Agreeing with 14 of the 15 appellate court assessments, the TS is unambiguous. Such general studies are inadequate for quantifying the effects in an individual case and do not comply with the standards for expert reports, which must formulate a reasonable thesis, be technically sound, and based on a verifiable dataset.11ECLI:ES:TS:2013:5819.
c) Judicial estimation after Tráficos Manuel Ferrer
A key decision taken by the TS concerns the questions whether the first and second instance judges were allowed to estimate the damage. The defendants argued that, because the claimants had provided only generic reports that were inadequate for quantification, they had simply not met their burden of proof. Furthermore, because the claimants also did not avail themselves of disclosure instruments, their inability to prove the damage is due to their own inactivity. In such a situation, the defendants argued, a judge cannot in effect heal the lack of evidentiary activity of the claimant by estimating a damage.
The Tribunal Supremo disagrees. In doing so, it rejects an extreme interpretation of Tráficos Manuel Ferrer according to which a court could estimate only if the claimant has at least tried to avail itself of disclosure instruments. The TS now clarifies that here the difficulty or impossibility to quantity the damage is not (primarily) due to the inaction of the claimant. The TS lists the reasons that make quantification exceedingly difficult in this case:
- There is hardly any suitable geographic comparator, because the cartel spanned the entire EU;
- A temporal comparison is difficult because of the long duration of the cartel;
- High complexity due to product heterogeneity;
- The settlement decision contains fewer details than a full decision;
- The pan-European nature of the cartel meant that many relevant documents are not in Spanish and hence possibly difficult for claimants to understand, should they demand them as part of disclosure;
- Many claimants reasonably assumed a pre-Directive law limitation period of just one year from the publication of the summary COM decision, leaving little time to conduct in-depth analyses (this assumption is not entirely correct, see below); and
- The cases about which the TS rules in this first wave stem from 2018, when the legal requirements were less clear and the factual assessment of the infringement across Europe was in its infancy.
Because of these difficulties, the TS concludes that the fact that the claimant relied on an inadequate expert report for quantification does NOT amount to evidentiary inactivity (“inactividad probatoria”), and neither does the fact that the claimant did not try to use disclosure instruments. In that discussion, the TS makes two interesting asides:
(i) The disclosure as currently practiced in Spain is not likely to be very helpful, and the work required would be disproportionate to the compensation sought; and
(ii) In the cases of the large claimants in front of the English Competition Appeal Tribunal,12Royal Mail Group Limited v DAF Trucks Limited and Others; BT Group PLC and Others v DAF Trucks Limited and Others, both 1284/5/7/18 (T),  CAT 6. where extensive disclosure was the basis for the reputable experts to conduct very costly econometric analysis, the court finally also rejected both sides’ analyses and estimated 5%.
Even if the claimant’s report is not good enough to be followed by the court, the court may estimate. Estimating an overcharge is not an illegal retroactive application of provisions of the Directive that stipulate presumption of damage, since Spanish procedural law leaves the door open for estimation in certain cases. If the conditions of article 386 of the Spanish Civil Procedural Law, Ley de Enjuiciamiento Civil (“LEC”) are fulfilled, the judge can, already under pre-Directive law, assume that there was some damage. Cartels are illegal precisely because they, usually, (“Id quod plerumque accidit”), lead to an artificial increase in prices.
The TS states that, given the characteristics of the cartel, the instance courts were justified to conclude that the trucks cartel was not among those “7%” of cartels (referencing a percentage from the Oxera report) that cause no overcharge. Defendants argue that 5% as a “conservative minimum” is arbitrary and that the courts could have set a lower percentage. The TS replies that the same could be stated for a higher percentage and that the defendants confuse arbitrariness with judicial discretion. It then confirms the application of a 5% overcharge in cases in which the claimant failed to prove a higher percentage based on a suitable expert report.
In the judgment in ECLI:ES:TS:2023:2475, the case in which the appellate court had upheld the award of 15% overcharge by the first instance court (based on the claimant’s generic report), the TS reduces the percentage to 5% (translation):
“These circumstances described in the Decision are also sufficient to understand that the damage was not insignificant or minor. However, it has not been proven that this damage exceeded the minimum foreseeable in a cartel of these characteristics, which has been prudently fixed by most courts at 5% of the cost of the trucks, since, as has been said, the lower courts have denied the evidentiary effectiveness of the plaintiff’s expert report, which fixed the damage at a higher percentage of overprice. As it has not been proven that the amount of the damage was higher than this minimum of 5% of the price of the truck, the exercise of the powers of estimation that the legal system attributed to the courts even before the implementation of the Directive, as a direct consequence of the principle of indemnity derived from arts. 1902 CC and 101 TFEU, does not allow them to fix a higher compensation. That is to say, as long as it is not proven that the amount of the damage has exceeded the minimum percentage of 5%, the plaintiff cannot claim compensation in excess of that percentage.”
The TS clarifies that legal interest is due from the time the damage occurred, meaning when payment was made. The reason for this is not delay (“mora”) but the principle of full compensation. The attacks levied against that by the defendants (arg: in illiquidis non fit mora) are hence beside the point. The TS cites the EU case law, starting with Marshall (1993)13ECLI:EU:C:1993:335. and the EU damage quantification guide14Commission Staff Working Document Practical Guide Quantifying harm in actions for damages based on breaches of Article 101 or 102. (para 24) on this point and confirms that this is acquis, which the Damages Directive15Directive 2014/104/EU. merely codified, as recently reiterated in Tráficos Manuel Ferrer.
On limitation, the TS rules that, while the five-year limitation period introduced by the Damages Directive cannot lead to a re-activation of claims that have become time barred, the new limitation period does apply to claims that are still “alive”, meaning, claims that had not yet become time-barred at the time the new rules came into effect or should have come into effect according to the deadline for the Member States to implement the Damages Directive. This is fully in line with the CJEU judgment in Volvo and DAF Trucks.16ECLI:EU:C:2022:494. The starting point (dies a quo) is the date of the publication of the summary of the Final Commission Decision in the Official Journal of the European Union on 6 April 2017.
4. Comment and outlook
The TS manages to balance various legal and policy concerns. In the absence of convincing quantification reports submitted by the parties, courts can look at the characteristics of an infringement to arrive at a conservative overcharge estimate. This approach is realistic in explicitly acknowledging that cartels are illegal precisely because they usually lead to higher prices. Allowing courts to estimate a conservative overcharge without prior disclosure proceedings enables “anyone” to bring a claim, without which the costs of the proceedings would far exceed the damages sought. A (mental) starting point even modestly above zero should, in the mid-term, help bring about earlier settlements and thus save judicial resources. At the same time, the decision leaves the door open for parties to bring high-quality evidence for quantification to convince the court of a higher (or lower) overcharge.
The evidence submitted and arguments raised can be different from case to case, so that it behoves the reader not to over-extrapolate based on this bundle of judgments. Yet, the TS does make clear that, in trucks cartel damages cases in which the lower instance courts found both the claimant’s and the defendant’s expert opinion unconvincing, the court should estimate an overcharge of 5% based on the characteristics of the infringement itself. This also applies to those cases in which the court fully followed the claimant’s expert report, even though that report was generic. Judging by the above pie chart, this outcome should apply to more than half of the well over 2,000 appellate judgements rendered thus far.
Yet, this will certainly not be the last time the TS will rule on damage quantification in trucks cases. It will be interesting to see how the TS will treat cases in which claimants did not just file a “generic” report, but collected a suitable dataset and conducted case-specific econometrics. Similarly, it is currently still open how the TS will rule in the roughly 6% of cases in which the appellate court awarded no damages, although, in at least some of those cases, the baseline approach of estimating a conservative overcharge of 5% should apply.
Beyond the trucks cases, the TS has made it clear that Spanish judges can estimate damages also under pre-Damages Directive law. However, whether purely “generic” damages reports will be enough to allow a judicial estimation in future cartel cases is not certain. Some of the reasons the TS lists for letting these generic reports suffice to avert the danger of being held “evidentiarily inactive” apply to most cartel damages cases. Other reasons, however, will not apply to future cartel damages cases, most importantly the argument that the presently decided cases were filed at a time when the requirements were less established. Furthermore, the specific amount of a 5% overcharge as a conservative baseline applies to the trucks cases. In cartels with different characteristics, the baseline might be higher or lower. It hence seems advisable, and, to go beyond the baseline indeed required, for claimants to provide case-specific econometrics based on a suitable dataset. For the construction of such a dataset, pooling of data will often be necessary to arrive at a set of sufficient variety and breadth. The assignment of claims to a special purpose vehicle which then brings one large, sophisticated action enables that without potentially illegal data sharing amongst competitors. That solution would also curb the flood of damages proceedings and would hence have efficiency advantages for the legal system.
On any cartel overcharge, legal interest accrues from the time the overcharge was paid. This is fully in line with long standing EU case law on this point. Similarly, regarding limitation, the TS confirmed what the CJEU established in case C-267/20 – Volvo and DAF Trucks: The five-year limitation period under the EU Directive on antitrust damages applies to all cartel damages claims that had not yet become time-barred when the Directive was implemented into national law.