Last week, the German Federal Court of Justice published its April 2021 judgment in Trucks II. The German Court – always good for a surprise – was overall less favourable to the claimant. In particular, it emphasised the role of economic party opinions and regression analysis, especially in relation to the factual presumption of price effects. Courts have to consider the data basis, methodology and result of the regression analysis when parties submit empirical party expert opinions.
Background
The case at hand concerns the famous Trucks cartel, which has kept competition litigators busy during the last couple of years. Between 1997 – 2011, six European trucks producers (Scania, MAN, Volvo/Renault, Daimler, Iveco, and DAF) coordinated gross list prices for medium and heavy trucks and the pricing and timing of technologies relating to the emission standards Euro 3-6 in the EEA. Following a leniency application of MAN, the case resulted in a hybrid settlement. In 2016, the Commission settled with MAN, Volvo/Renault, Daimler, Iveco, and DAF, which were fined € 2.93 billion in total. Scania, who decided not to settle, was fined € 880 million in 2017 for participating in the trucks cartel.
Following the fining decisions, many direct and indirect purchasers claimed damages all over the EEA. Naturally, also the German courts have to deal with numerous claims, which lead to considerable differences between the decisions of lower and higher courts. Already last year, the German Federal Court of Justice ruled on many of the complex issues in Trucks I, which our esteemed author Thomas Thiede discussed for the Kluwer Competition Law Blog. In April 2021, the Court followed with Trucks II – another puzzle piece for German private enforcement of competition law.
On the sidelines…
Before turning to the most compelling matters, the Federal Court of Justice also had to put out a few “smaller” fires.
Once again, the Court emphasises the two-limb causality test under German law: the differences between the establishment of liability in principle (Kartellbetroffenheit) and the cartel affectedness (Kartellbefangenheit). What has initially been set out in the Rails II judgment more and more becomes established case law.
Even though it falls outside of its competence, the Federal Court of Justice presents quite lengthy instructions on the passing-on defence. In German law, the passing on defence is captured by the general legal instrument of benefit set-of (Vorteilsausgleichung). Naturally, the Federal Court of Justice applies the standard rules and limitations of this legal instrument also in private enforcement of competition law. In the present case, the Court underlines the possibility to exclude the passing-on defence altogether based on normative grounds. The damaging event must not unreasonably burden the injured party and, in particular, must not favour the damaging party unfairly. In that context, the Court emphasises the role of private enforcement as an integral part of the effective enforcement of competition law and its role in guaranteeing the public interest of undistorted competition. In particular, the passing-on defence can be excluded if the (large number of) indirect customers are unlikely to claim damages, for example, in the case of scattered damages where only a relatively minor claim can be considered for the individual indirectly injured party. With the Damages Directive giving a right to passing-on defence and the context of the normative exclusion relating to the public interest of effective competition law enforcement, a clarification by the European Court of Justice almost certainly will lie in the future.
The judgment also contains a long section on the suspension of the statute of limitations. Without going into detail here, the Federal Court of Justice clarified that the suspension period does not begin with the notification of the fining decision but with the expiry of the period for filing the action for annulment pursuant to Article 263(4) TFEU.
But, let’s once again turn to price effects and damages causation, where the German Federal Court of Justice discussed two pertinent issues…
Factual presumption of price effects for coordination of gross list prices
From the various Rail judgments, we know that there is no prima facie evidence (Anscheinsbeweis) that cartels result in higher prices and, therefore, damage. Instead, according to the Federal Court of Justice, a factual presumption (tatsächliche Vermutung) of price effects may apply, which carefully needs to be weighed against other available indicators – more on the latter in a minute.
In Trucks II, the Federal Court of Justice clarified what was already indicated in Trucks I: the factual presumption also applies even if the cartelists only agreed on gross list prices and not on the prices to be paid by the end customers. The Court held that the list prices typically formed the starting point for pricing discussions, which necessarily meant that they affected – in some way – the transaction prices to be paid by the customers. The Court admits that market pricing depends on numerous factors. It held that the list price, which is typically significantly higher than the transaction prices paid by the customers, is only one of these factors. However, in the view of the Court, this only indicated that the relationship between list and market price is variable and that there is no “systematic” or fixed connection. However, it does not prevent the assumption that the cartel agreement adversely affected the transaction prices achieved on the market with a high probability.
Interestingly, the Federal Court of Justice mentioned the binding effect of the Commission decision several times in the context of price effects. Even though the German Court does not refer to the relevant provisions concerning the de jure binding effect (Article 16 Regulation 1/2003), it held that the final decision of the European Commission is binding when it comes to the anti-competitive conduct, the coordination of gross list prices, also in the context of assessing the price effects. Furthermore, it focused on the binding findings of the Commission decision, which found that the list prices set by the respective head office typically formed the starting point for pricing discussions. Normally, price effects, the damage itself is – naturally, as the competition authorities do not calculate damages – not part of the de jure binding effect of competition authority decisions. However, the Trucks II judgment shows us that at least a de facto binding effect might influence the damage assessment (German speakers should read Becker’s paper on de facto binding effects).
Moreover, the Federal Court of Justice also applied the factual presumption of price effects in cases where the new vehicles were not purchased directly from the manufacturer but from a legally independent dealer. Here, the Court came up with something new – it did not have to take recourse to the ORWI case-law according to which pass on of costs is likely if most of the customers acting as suppliers on the next market level have to pay the cartel price and competition on the aftermarket is otherwise functional. Surprisingly, it held that a clear separation of different market levels does not exist in the context of the Trucks cartel at all. The legally independent truck dealers do not represent a continuous market level between the manufacturer and the final purchaser. Instead, according to the Court, they are integrated into the manufacturers’ distribution structure, who themselves distribute their products partly directly or indirectly via dependent dealers. Also here, the list prices used at the wholesale level form the basis for the price agreement, in which the given leeway for granting discounts – either to the dealer or directly to the end customer – could be used. While the latter statement holds up, merging the market levels is a bit of a stretch in terms of reasoning. Even integration into the retailer’s structure does not make market levels disappear.
With its judgment, the Federal Court of Justice reinforced the factual presumption once again. Yet, the Damages Directive and 10th Amendment of the German Competition Act will provide for further-going legal presumptions. Generally, with strengthening the factual presumption, the Court appears to be claimant-friendly here. However, this is immediately called into question by the Courts focus on the weighting of indicating factors, of which the factual presumption is just one of the indicators. More on this now!
Factual presumption v. weighing of indicating factors: the role of economic party opinions and regression analysis
As mentioned above, according to the German Federal Court of Justice, the factual presumption of price effects carefully needs to be weighed against other available indicators (Gesamtwürdigung). In Trucks II, the Court again underlined the importance of the overall weighing assessment in relation to the factual presumption. The factual presumption should not be given undue weight. In fact, the Court held that the weight of the principle of experience depends on the cartel’s concrete structure, duration, stability, etc.
When considering the overall context of the infringement, the Court also relied on the Commission decision’s binding effect. Underlining its findings in Trucks I, the Court of Justice held that the limits of the binding effect must be observed. However, further conclusions can be drawn from the binding findings of the Commission decision. Non-binding findings can be appropriated as long as they have not been contested or not sufficiently contested by the cartel participants. Thus, the Federal Court of Justice does not rely on the de jure binding effect as intended by Article 16 Regulation 1/2003 (or Article 9 Damages Directive). Instead, the Court once again takes recourse to a de facto binding effect.
In the present case, the Federal Court of Justice held that the Court of Appeal wrongfully conducted the weighing exercise since it did not consider the regression analysis of several economic party opinions. For the damages estimation standards of § 287 Code of Civil Procedure, a clearly predominant probability based on a sound foundation that damage has occurred is sufficient. However, any circumstance of indicative value must be included in the assessment.
First, the German Federal Court of Justice followed the preceding instance and held that parts of the party expert opinions indeed incorrectly assumed that there had only been an exchange of information between the cartelists as the basis of the respective economic analysis. However, the Court then held that the defendants and interveners expert opinions also contained econometric regression analysis of transaction prices during and after the cartel period. This analysis determined whether a systematic difference in transaction prices could be identified between the two periods. Therefore, the party expert opinions also contain a comparative market analysis independent of the type and nature of the concrete agreements. Yet, the expert party opinions were entirely ignored by the court of appeal.
According to the Federal Court of Justice, such expert party opinions must be observed. As a reminder, expert opinions can take two different forms in cartel damages actions like in any other German civil proceedings: court-appointed expert opinions or party opinions. The court-appointed expert is obliged to the court and supports it if the court lacks expertise; nevertheless, the court remains free in assessing the evidence and is also not bound by the expert opinion. Party expert opinions do not constitute evidence but qualified party submissions. Yet, the Federal Court of Justice has now considerably upgraded the value of party expert opinions for the weighing exercise regarding cartel damages liability, especially in relation to the factual presumption. In particular, regression analyses need to seriously be assessed and taken into account in the weighing exercise. When empirical party expert opinions are submitted, the Federal Court requires a consideration of the data basis, methodology and result of the regression analysis.
In the present case, the respective regression analysis submitted by the intervener and the defendant showed no economic evidence for a deviation of the transaction prices paid during the cartel period from the hypothetical market price. A different outcome of the weighing exercise seems, therefore, likely. However, the Federal Court of Justice also held that when retaking the overall weighing of all indications relevant for the determination of the occurrence of damage, the court of appeal would need to examine – if necessary with court-appointed expert support – the resilience of the regression analyses submitted by the parties. Nevertheless, the Federal Court of Justice also underlined that any court is still entitled to considerable methodological leeway within the scope of the power of estimation. This means that the methods chosen by the parties are not set in stone. A court might recourse to another method or other comparative data, as long as it thereby meets the specified objective of coming as close as possible to reality using probability considerations with an effort appropriate to the matter.
In this context, the Federal Court of Justice, along with its reasoning in Rails II, further questions the use of interlocutory judgments (Grundurteil). The Federal Court of Justice referred the proceedings back to the court of appeal. It recommended that instead of deciding once again only on liability, the procedural economic decision might be to turn directly to the determination of the amount of damages.
Conclusion
In my opinion, the judgment stands out for two issues.
First of all, it shows the vast impact of the fining decision on private enforcement of competition law and damages estimation. The fining decision has a binding or at least indicative, de facto effect in the damages assessment and is an essential indicator in the weighing exercise. Thus, fining decisions and their content should be analysed in-depth and harnessed for private enforcement purposes.
Second, and most of all, the decision clearly shows the important role of economic party opinions and economic expertise in private enforcement of competition law in general. This will undoubtedly continue the trend of providing ever-lengthening divergent party opinions (and increasing costs for parties) in cartel damages actions. Once again, the Federal Court of Justice does not exactly take a claimant-friendly view. Rather, cartel damages actions will remain a lengthy and costly issue. The judgment emphasises the necessity for courts to deal with highly complex economic expert opinions in greater depth. Since the courts’ own expertise in this respect leaves much to be desired, court-appointed experts have to save the day. A court-appointed expert will be able to assess whether the regression analysis is correct in itself. However, it will probably not be able to evaluate which different underlying data is correct. This naturally creates problems for the final damages calculation.
Luckily, the Federal Court of Justice gives lower courts considerable methodological leeway when estimating the damages. The involvement of economic expertise is nevertheless necessary, also when considering other methods. Reliable economic expertise and established methods remain key in private damages actions. This will also not change much with the introduction of the presumption of harm introduced by the Damages Directive and supplemented by the 10th Amendment of the German Competition Act. The legal presumptions only cover the occurrence of harm, not the amount. The calculation of the amount of damage will largely depend on economic expert opinions (as previously discussed here).
________________________
To make sure you do not miss out on regular updates from the Kluwer Competition Law Blog, please subscribe here.
I am afraid, I have to disagree on three matters:
1)
The Federal Court of Justice – just as many German lower courts – quite correctly ‘merged’ the manufacturers distribution structure as we had quite some experience with an avoidance of liability through restructuring (as we call it ‘sausage gap’). And, I think with a view to CJEU’s judgments in Otis and Skanska this is quite correct: If it says, for the lack of a better example, Hornkohl on the outside of the shop, there is Hornkohl inside the shop, isn’t it?
2)
Indeed the Federal Court of Justice addressed the question of assessing the regression analysis but asked only for taking it into account as one (of many) aspects and, to be sure, asked for the assessement of the underlying data.
Be that as it may, quite notably the Federal Court of Justice held in no 88:
“In view of this, the court of appeal will also … have to consider … to turn directly to the determination of the hypothetical market price and thus to the reason for and the amount of any damage suffered by the plaintiff. Within the scope of the power of estimation to which it is entitled in this respect, the trial judge and thus also the court of appeal are entitled to considerable methodological leeway, so that it may also use a method other than the one chosen by the parties and other comparative data, as long as it thereby meets the specified objective of coming as close as possible to reality by means of probability considerations with an effort appropriate to the matter.”
This means, in essence, to freely estimate the damage in any way the court wants as long as it comes close to reality – and exactly no need for further or indepth emprical expert opinions.
3)
I regularly wonder – and I think this is striking in the assessment of expert opinions in the Truck cartel – how no economic evidence for a deviation of the transaction prices paid during the cartel period from the hypothetical market price could be established at all. I indeed get the math but, alas, the real world doesn’t fit: Why would the Commission fine the manufacturers if this had no impact at all? Why did the manufacturers uphold this practice when it was of no economical use? Why risk considerable fines or even prosecution?
Dear Thomas,
thanks for your very extensive comments! A pleasure to discuss with you.
On 1): I understand what you mean re Otis and Skanska but is an integration in or rather the usage of the manufacturers’ distribution structure by otherwise independent dealers sufficient to fulfil the Otis and Skanska criteria? Yet, we all know that the CJEU also likes to expand those. Guess we have to wait for further case law from Luxembourg, also in the context of the Damages Directive.
On 2): I agree and have stated above that the BGH puts the focus on regression analysis in the sense that if a party opinion contains such an analysis, the lower court has to sufficiently deal with it. Yet, it still gives methodological leeway to lower courts, i.e. they can choose other credible (economic) methods for damages calculation. Since courts are lacking economic expertise to assess these methods, in my opinion, court-appointed economic experts are still necessary. If the free estimation method that does not need economic expertise is such a valuable method… I think we will disagree (or you can try to convince me on Friday at the CLF event)!
On 3): I must say, I am quite torn. On the one hand, I am a big fan of sound economic expertise in cartel damages actions (in various forms). On the other hand, knowing economists, there are always ways to bring the amount of damage down to zero, so economic party opinions should still be taken with caution.
I guess there is still a lot of issues to discuss!
You make some excellent points, TT. If I may, I’d like to offer some answers to the rhetorical questions you ask in 3):
“Why would the Commission fine the manufacturers if this had no impact at all?” Possibly because the European Commission, under Margrethe Vestager, was seeking to send a strong signal to European industries that exchanges of information, which may have been tolerated in the past, were no longer going to be overlooked. An implementation of a policy position, in other words. When MAN sought immunity as the whistle-blower, it presented the Commission with the ideal opportunity to send this strong signal by making an example of these truck makers.
“Why did the manufacturers uphold this practice when it was of no economical use?” It may have been of economical use without having an impact on transaction prices paid during the cartel period. In those years, the European truck manufacturers were anticipating an enormous oncoming wave of South Korean, Chinese and US imports, three countries that benefit from high levels of corporate cooperation/collusion and state support, and little antitrust activism. In the face of such challenges, the Europeans exchanges of information were an attempt to prepare themselves for future market conditions.
“Why risk considerable fines or even prosecution?” Previous European Commission antitrust practice around exchanges of information had been relatively benign. The Trucks fines were at the time outliers – the largest fines to date, in fact. One must look at the risk as it would have appeared to the participants one-and-a-half decades ago.
In other words: while I agree that counter-intuitive results from econometric analyses naturally inspire scepticism, they are just that – counter-intuitive. “We’re entitled to our own opinions, but not our own facts”. There are many possible explanations why companies, acting without culpable intent, might have engaged in an exchange of information that did, in the end, truly have no impact on the prices paid.