Amsterdam Court rules on how to calculate damages in air cargo cartel cases

On 6 November 2024, the Amsterdam District Court ruled on the methodology to be used for the damage calculation in follow-on claims of indirect purchasers from the air cargo cartel. The court confirmed that a regression analysis is an appropriate method to calculate damages and ruled that in this case an analysis on the level of the indirect purchasers should be used (the “one step model”) instead of an analysis on the level of the direct purchasers (the “two step model”). 

Background of the case

The case concerns follow-on damages claims by two claim vehicles on the basis of the air cargo cartel. This air cargo cartel was established by the European Commission and concerned illegal coordination of fuel and security surcharges for air cargo services by airlines.

The claim vehicles brought their case on behalf of indirect purchasers of air cargo (the “shippers”), that had assigned their claims to the claim vehicles. The shippers had purchased air cargo services from freight forwarders, who had in turn purchased them from the airlines. The shippers claimed that they had paid inflated prices for the services as a result of the cartel.

In one of the earlier interim decisions in this case, the court had determined for which shippers the claim vehicles had adequately substantiated that they had purchased at least one air cargo service during the cartel period — and had therefore shown that they had likely suffered harm. 

In this decision, the court addressed the methodology for calculating damages. The court determined the appropriate economic method to be used by the experts, at which level of the distribution chain the analysis should be conducted and what data should be used.

The discussion between the economic experts

The court had requested the economic experts of the parties to prepare a joint “agree/disagree” document, outlining the points of agreement and disagreement concerning the methodology for determining the overcharge and the data to be used in the overcharge analysis. The economic experts agreed on a number of points.

First, they agreed that regression analysis  (a statistical technique to analyze the relationship between economic variables) is the appropriate method to determine whether an overcharge occurred. They also concurred on the need to account for factors unrelated to the cartel’s impact on air cargo service prices by incorporating control variables in their analysis, such as the weight of the cargo, fuel costs, and fluctuations in demand for air cargo services.

Second, they agreed that air cargo prices during the cartel period should be compared to prices after the cartel (a during-and-after comparison). Prices from before the cartel period would be excluded from the analysis due to insufficient data for a robust assessment.

Third, the experts agreed that the overcharge analysis should take account of (i) to what extent the overcharge resulting from the introduction, coordination, and application of fuel and security surcharges led to corresponding reductions at the total price level – a concept known as the ‘waterbed effect,’ where price increases in one area (such as surcharges) are partly offset by price reductions in other areas; and (ii) to what extent freight forwarders passed on the overcharge to the shippers (the upstream pass-on).

The economic experts of the parties disagreed whether to use the airlines’ data or the shippers’ data.

The airlines argued that their data was more suitable because it (i) was the most direct source material, (ii) included surcharges (which directly reflects the illegal agreements between the airlines); (iii) had more data points; and (iv) avoided contamination from external factors like freight forwarder costs and a separate freight forwarder cartel. The airlines had not yet submitted their data, and only offered to make the data available during an oral hearing about this topic.

The claimants argued that their data should be used. Although smaller, their dataset provided enough high-quality data to allow for a robust statistical analysis. Additionally, the shipper dataset was already cleaned and had already been provided to the defendants, whereas the airline data would still require an extensive and time-consuming matching process between the airline and shipper data. The claimants also noted the potential bias in the airline data, as not all routes and airlines were included in the dataset.

Secondly, and tied to the first dispute point, the parties disputed on whether to use a one step model or a two step model. This relates to the question whether the overcharge should first be calculated at the level of sales by airlines to freight forwarders, followed by an assessment whether freight forwarders passed on any overcharge to the shippers (the two step model), or whether only the transaction data of the shippers could be used (the one step model).

The airlines’ main objection to using the one-step model was that it offers no insight into the pricing behavior, and in particular the margins, of the freight forwarders, since it only uses shipper data.

The claimants argued that it is unnecessary to separate the freight forwarders’ margin from the total price. Their economists found no evidence that the variation in freight forwarders’ margins was linked to the cartel period, as there was no significant difference in how they priced their services before and after the cartel. Therefore, the freight forwarders’ margin did not affect the overcharge analysis in the one-step model. Furthermore, the “black box” issue regarding the freight forwarders’ price setting also existed in the two-step model, as the airlines had no more information about this than the shippers.

The airlines and shippers also debated whether the overcharge analysis should focus on the surcharges alone (airlines) or the total price, which includes the base price, surcharges, and freight forwarder margins (shippers).

The claimants argued that analyzing the total price accounts for the “waterbed effect,” where higher surcharges could lead to reductions in other price components, balancing the overall price. They also pointed out that analyzing only surcharges is problematic because (i) there are no benchmarks for surcharges before the cartel; (ii) surcharges do not represent a separate service; and (iii) it is often impossible to isolate surcharges from the total price in the data, particularly the shipper data.

The airlines questioned the reliability of the claimants’ approach, arguing that focusing on total prices makes identifying overcharges harder. They also found it implausible for surcharges (20% of the total price) to cause a 10% overcharge on the total price, as this would imply the overcharge was nearly as large as the entire surcharge itself. The claimants’ economists responded that it does not render the explanation implausible. In fact, they noted that the entire surcharge could indeed represent the total overcharge. Before the cartel, no surcharges existed, so introducing a surcharge of 20% would constitute an overcharge of 20%. The difference between the total surcharge and the shippers’ estimated overcharge on the final price could potentially be explained by the waterbed effect (a lowering of the base price).

The court’s decision

The court sided with the claimants on all points, ruling that the overcharge should be calculated using the shipper data through the one-step model applied to the total price paid by the shippers. The court found no reason to doubt the claimants’ proposed method for conducting a reliable statistical analysis.

The judgment confirms that there is no single best method for damages calculation: both the one step model and the two step model can be suitable for calculating overcharges in cartel cases. The appropriate approach depends on the specifics of the case and the available data. In this instance, the data of the shippers were already accessible, and the court agreed with the drawbacks of using the airlines’ data as pointed out by the claimants.

Observations

The decision is interesting because it provides guidance how Dutch courts deal with damages calculations by economic experts in antitrust cases. It is, as far as we are aware, the first time in a cartel damages case that the court has instructed parties’ experts to prepare an agree/disagree document to outline their differences. The economic experts also debated their differences during a court hearing. The court’s approach seems to have worked well to focus the discussion. The judgment confirms, with reference to the European Commission’s Practical Guide for the calculation of harm, that there is no single best method to calculate damages: both the one-step model and the two-step model are suitable for calculating overcharges in cartel cases. The appropriate approach depends on the specifics of the case and the available data. In this instance, one of the factors in the court’s decision seems to have been that the claimants had already shared their data, while the defendants only offered to do so at the hearing. A lesson to learn from this is that it does not always pay off to keep your cards close to the chest.