The General Court has upheld an appeal against the European Commission’s UPS/TNT merger prohibition, over four years after the Commission issued its decision. It concluded that the Commission should have made the final version of its model available for the parties to review. Although there is some risk that this case could make competition authorities think twice about relying on quantitative evidence, it also represents an opportunity for them to become more transparent, albeit within the constraints that they face.
Background to the UPS/TNT merger
In March 2012, UPS announced its plan to acquire TNT. Both companies were major players in the small-parcel delivery market in Europe. The Commission started investigating in June 2012, and soon raised concerns over price rises that were likely to result from the merger.
The Commission and the merging parties conducted extensive economic analysis to assess whether prices increase with concentration in express parcel delivery markets in the European Economic Area (EEA). The main econometric model used was a price concentration analysis (PCA). PCA aims to estimate the impact on prices of a reduction in the number of competitors.
The PCA in this case assessed the effect of a reduction in the number of parcel operators on a ‘lane’ (an origin–destination pair). The merging parties and the Commission agreed that the analysis showed that parcel delivery prices do increase with higher levels of market concentration, but disagreed on the magnitude of the effect.
After a detailed phase 2 investigation, the Commission concluded in January 2013 that the merger would be likely to result in price increases in 15 EEA countries where countervailing factors (such as merger-specific efficiency gains) were not significant enough to offset the loss of competition. The Commission deemed that the merger would therefore lead to a significant impediment to effective competition (SIEC) in these countries, and blocked it. The merging parties appealed the Commission’s decision. (Meanwhile, TNT has been taken over by another rival, FedEx, a deal approved by the Commission in January 2016.)
The General Court judgment: bad news in the post for the Commission
In March 2017, the General Court aned the Commission’s decision on the basis of the right of defence of parties. In its judgment, the Court noted that:
Accordingly, the applicant’s rights of defence were infringed, with the result that the contested decision should be aned, provided that it has been sufficiently demonstrated by the applicant not that, in the absence of that procedural irregularity, the contested decision would have been different in content, but that there was even a slight chance that it would have been better able to defend itself [paragraph 210]
The Court found that it was ‘clear from the documents in the file that the final version of the econometric model was not communicated to the applicant’. The failing that the Court identified is indeed a fairly technical point, but it is non-negligible:
the Commission relied on two different variables at the stage of the statistical estimation of the effects of the loss of a competitor on prices and at the stage of the prediction of the effects of the merger on prices. [paragraph 206]
Thus, the Commission relied on a discrete variable at the estimation stage and on a continuous variable at the prediction stage. [paragraph 207]
Although the use of a discrete variable had been discussed repeatedly during the administrative procedure, it does not appear from the file that that was also the case as regards the use of different variables at the different stages of the econometric analysis. [paragraph 208]
In statistics, a continuous variable can take on any value between two specified values, whereas a discrete variable can take only certain values from a finite set. For instance, the number of competitors in a given market is a discrete variable, as it can be specified only in whole numbers; on the other hand, the market share of firms can be expressed as a continuous variable (e.g. 40.19%).
The judgment of the Court in UPS/TNT, while focused on a point of law, at the same time illustrates the critical importance of transparency in the use of economic evidence.
Transparency cuts both ways
Economic evidence often plays a key role in competition investigations, and parties in competition cases are asked to follow the competition authorities’ guidance on the submission of this evidence, as described in the box.
Competition authorities’ guidance to parties
Economic evidence needs to be presented in such a way that it allows for a proper review by the economic experts at the competition authority and, in some cases, by third parties. Various guidance documents on best practice for economic submissions have been issued by competition authorities. These follow some common principles, which are articulated as follows in the UK guidance.
Clarity and transparency—submissions should not only present the results and conclusions of the economic analysis undertaken, but they should also clearly state the methodology used, the assumptions made in reaching the results, the justification for the methodology and the assumptions, and the robustness of the results to any assumptions made.
Completeness—submissions should contain a complete description of the analysis undertaken. All relevant assumptions should be discussed and the choice of techniques explained. Relevant econometric output, diagnostic tests and checks for robustness should be presented.
Replication of results—the competition authority may want to replicate the results of the analysis that has been submitted. This means that parties should be prepared to respond to a competition authority’s request, at very short notice, for all relevant computer code and data files necessary for the authority’s economists to reproduce the results.1
The European Commission’s guidance contains some useful additional points on presentation and interpretation of the results.2
Source: 1 Competition Commission (2009), ‘Suggested best practice for submission of technical economic analysis from parties to the Competition Commission’, 24 February, pp. 1–2. 2 European Commission (2010), ‘Best practices for the submission of economic evidence and data collection in cases concerning the application of Articles 101 and 102 TFEU and in merger cases’, January.
Transparency of competition authorities: clear benefits
As the UPS/TNT appeal has shown, competition authorities also have a duty to share their economic evidence with parties under investigation—particularly the final versions of any models that are used to inform overall case decisions such as whether to clear or block a merger.
This is not the first competition authority decision to be overturned due to a lack of transparency with regard to economic evidence. In December 2014, the UK’s Competition Appeal Tribunal (CAT) aned part of the final report into the private healthcare market investigation by the Competition and Markets Authority (CMA). One reason was that the CMA made material errors in its price analysis. Some of these errors came to light only after the CAT ordered the CMA to grant the parties additional access to its econometric analysis via a data room. The CMA acknowledged that it had made a procedural error by not consulting the parties after altering the analysis. Transparency is important for complex models, as technical assumptions can have an impact on the results; even if the parties only audit the authority’s model and do not submit their own, this is likely to be a useful check on the quality of the authorities’ work.
Data rooms and confidentiality rings
Transparency has improved over the last decade. The European Commission published its guidelines on best practice in January 2010 (see the box above). Data rooms and confidentiality rings are now more frequently permitted by competition authorities for parties to examine economic evidence and models.
However, it is also important to acknowledge that competition authorities face significant time constraints in merger proceedings, and that they have to ensure confidentiality. These issues create challenges and limitations, even for competition authorities wishing to maximise transparency. The data room process often involves a significant amount of time and effort for the authority and for external lawyers and economists; for example, strict confidentiality rules mean that parties’ advisers cannot take any written material out of the data room without the authority’s approval, and they are not allowed access to the Internet or to phones and other wireless devices. All sessions must be supervised by officials, and thus have limited working hours.
Nevertheless, competition authorities can plan ahead for potential data exchanges, confidentiality rings and data rooms in the course of the investigation. This can reduce the risk of decisions being overturned due to a lack of transparency or modelling errors.
It remains to be seen whether the Commission will appeal the UPS/TNT judgment to the Court of Justice. However, if the General Court’s judgment stands, this should be welcomed as it demonstrates that courts can provide guidance to regulators and competition authorities on transparency when it comes to the use of economic analysis in decisions.