The Commission published the text of its most recent prohibition decision in Deutsche Boerse / NYSE Euronext. The Decision is lengthy and the Commission appears to have formulated a response to most arguments proffered by the parties.
However, a review of the Decision brings to the fore a number of ways in which the Commission could improve the quality of the evidence which it uses to support its decisions. Many of the issues discussed below have been debated in the past, including before the EU Courts. And while the Courts generally show a certain level of deference to the type of evidence which the Commission uses, that should not prevent the Commission from continuously trying to improve the quality of its decisions and the underlying evidence.
The first issue is the lack of quantification. The DB/Euronext Decision explains that the parties criticized the Commission for not conducting any empirical, economic or econometric studies, in particular as concerns market definition in the area of derivatives. The Decision responds that the Court in its 2010 Ryanair/Aer Lingus judgment, concluded that quantitative analysis could be useful but is not mandatory and that there is no hierarchy between different types of evidence (i.e., quantitative versus non-quantitative). However, regardless of the Court’s conclusion, it is undeniable that robust quantitative evidence can be particularly convincing and that as a general matter, the Commission should strive to quantify its analyses as much as possible (similar to what notifying parties are asked to do when submitting an efficiency defence).
The Commission’s second argument is that the necessary data were either not available or not of adequate quality. The Commission adds, specifically with regard to the issue of quantification of the level of competition between exchange-traded derivatives and over-the-counter (OTC) products, that “data on OTC trading are very difficult to obtain due to the very nature of OTC market” (sic) and cites the parties themselves saying that “[i]t is generally very challenging to produce exact figures for the opaque OTC segment.” Last, the Commission justifies the lack of quantification on the basis of the fact that the parties themselves were not able to perform any such analyses in relation to market definition.
The fact that data are difficult to obtain and that notifying parties themselves have not been able to provide an analysis are rather poor excuses. The EUMR provides very extensive powers of investigation to the Commission, including to obtain data from third parties, and data to which the notifying parties by definition do not have access. The Commission should use these powers to obtain data in cases where they are needed for a robust quantitative analysis.
The Decision also does not provide quantification in terms of economic harm. Its conclusion on derivatives is simply that the Parties are each other’s closest actual and potential competitors and that because of the elimination of this significant competitive constraint, and the creation of a merger to near-monopoly, the choice of platforms for derivatives trading will be significantly reduced “which is likely to lead to higher exchange fees and less innovation.” But the Decision remains silent on the level of the predicted price increases.
The Decision argues that the lack of quantification is compensated by the two other evidentiary tools which the Commission generally uses in its investigations: an extensive market test and an analysis of the parties’ internal documents.
Given the pivotal role of both types of evidence, the Commission should also here ask the question how they can be improved.
The market test in the DB/Euronext Decision appears impressive: in Phase I, 600 detailed requests were sent out covering seven different groups of market participants and more than 250 responses were received and analysed. In Phase II, the Commission sent 150 questionnaires to targeted marketed participants active in the most relevant markets and received over 100 responses. The Commission also conducted 20 teleconferences and meetings with customers and competitors.
But are these market tests executed correctly? The Ryanair judgment (in para. 214 and following) is very generous on the review of Commission market tests: “The Commission can thus not be accused of having acted incoherently or unreasonably on the sole ground that it attached less importance to the responses which it considered to be less relevant.” But the organization of surveys is an art or a science (or maybe both) and the question is whether the Commission sufficiently mastered it. What is the correct sample and how should the respondents be selected? What is the correct way to formulate questions and avoid bias or leading questions? How many questions should be included in a questionnaire – many companies in the EU have experienced the joys of 150-200 question information requests? How should a mere count of the number of responses be weighed against the verbatims in the responses? How should the more relevant third parties or responses be distinguished from those that can be dismissed and how can bias in the responses of third parties be detected and interpreted correctly? Last, the Commission questionnaires often ask very complex questions and it needs to consider to what extent respondents have the knowledge and sophistication to respond or at least whether third parties have taken the time and effort to analyse the complex issues or their responses are merely “cheap talk”.
In other words, does the Commission have the necessary know-how in-house to conduct these surveys, which are an essential part of its decisions, in a correct and professional way?
The third pillar of the Commission’s evidence gathering is the review of internal documents. The main criticism often heard is that the Commission tends to cherry-pick: to select those documents and quotes which appear to support a particular point it wants to make. In other words, reviewing Commission decisions often leaves one with the impression that the Commission formulates a specific point and then investigates whether internal documents contain language that can support this point.
In DB/Euronext, it appears that the Commission has used a number of ordinary course documents or white papers published by the parties and third parties, as well as reports prepared by third-party market analysts. However, the Decision is also replete with selected quotes taken from an “NYX internal document” or a “DB internal document.” It is not apparent from the text of the Decision that the Commission has carefully reviewed all internal documents of both parties for information on a given argument and carefully weighed the evidence for and against the argument. An equally serious issue is that one “internal document” is not equal to another: in large organizations, hundreds and thousands of internal documents and correspondence are created but not all of them carry the same weight within the organization. The Commission should consider that there is a hierarchy between different documents and different authors and make it apparent in its decisions that it takes these hierarchies into account when weighing the evidence taken from internal documents. On any given topic, should minutes of a Board meeting or reports prepared for a Board by senior management be given more weight than a note or e-mail prepared by a local sales manager? At least from the text of the Decision, it is not clear how or whether the Commission has done such necessary weighing exercise.
The Ryanair judgment (rightly or wrongly) gives the Commission significant freedom in the selection of its evidence. However, the Commission should take this freedom as a responsibility and continue to work hard to improve the quality of the evidence used in its enforcement decisions.