Infringement decisions adopted by the European Commission in competition procedures are subject to review by the Court of Justice of the European Union (‘CJEU’). The CJEU is an international tribunal comprising judges from countries with varying legal traditions within Europe.
This empirical paper, available here, seeks to determine whether the outcome of appeals from European Commission infringement decisions are affected by the legal origins of the judge rapporteur. In particular, the authors test whether judges coming from countries with French-inspired administrative systems are more likely to decide in favour of the Commission. They find that this is indeed the case, and that the results are robust to alternative political ideology variables, including left–right politics and a preference for European integration. The paper is structured as follows:
Section 2 delves into the various legal traditions of Europe, and how they may influence the judicial review of competition appeals.
There is extensive literature demonstrating that preconceptions nourished by education, work experience and political ideology, among other possible determinants, can have a profound impact on how judges evaluate facts and reach a decision. When judges are confronted with complex facts and circumstances, and cannot reach a conclusion through pure and straightforward legal reasoning, their intuition, which is often influenced by legal training, is likely to come into play.
In addition to the traditional distinction between common and civil law, the EU is home to three main civil law families: French, German, and Nordic. The most common legal tradition in Europe is the French one – a tradition which grants a larger role for the state and greater deference to the administrative branch, and relegates the judiciary to a bureaucratic and subordinate role. Substantively, French administrative law has been observed to intrude less in decisions made by the executive, which is entrusted with the freedom to pursue the collective public interest, than other jurisdictions. This is also reflected in the institutional arrangements for reviewing executive action. While the common law does not distinguish between public and private law and offers the same protection to all legal actors, the French tradition offers greater protection to the state. Procedurally, administrative cases in the French system are handled by specialised administrative courts and judges are drawn from a distinct corps of the administration.
The French model has been followed by most other Continental European countries, with the authors noting that French administrative law not only has a deep influence on countries with French legal origin, but also on the German law tradition and in Eastern European countries. On the other hand, the French model has a relatively weak influence on Nordic countries – where Denmark and Norway do not have administrative courts and Sweden’s courts have more extensive powers of review than elsewhere in continental Europe.
Section 3 investigates the CJEU’s decision-making process and analyses how judicial panel dynamics might affect how individual judges vote on a case.
Competition appeals in the EU are heard at first instance by the General Court, and by the CJEU on points of law only at second instance. The General Court – the subject of this study – mainly deals with fact-intensive cases involving competition, state-aid, trade, agriculture and trademarks.
EU judges work together in a chamber (usually of three, exceptionally of five judges). Before the 2015 reforms, all incoming cases were allocated to chambers in turn following three separate rotas, one of which was for competition and state-aid cases. After a case is allocated to a chamber, its president will assign the case to a judge as rapporteur. Accordingly, case allocation to chambers largely follows an automatic process, while the designation of the rapporteur has elements of discretion since a chamber president may consider expertise, workload and other policy factors that are very individualised when choosing the rapporteur. The General Court issues a single judgment on behalf of all the judges, and does not allow for dissenting opinions. As a result, the actual votes of individual judges are not observable.
In theory, each judge sitting in a panel is entitled to the same voting power. However, the influence of individual judges may vary in practice depending on their particular roles in a case. The rapporteur assumes the responsibility of drafting the report of the hearing, which is essentially a summary of the parties’ arguments, and a preliminary report, an internal document for purposes of deliberation that summarises the legal and factual background of the case and contains the personal observations and recommendations of the rapporteur. This report is then circulated to the other judges. After deliberation, the rapporteur incorporates any comments from the other judges, with the aim of ensuring that a consensus is reached among panel members. Given this, the rapporteur is the judge better placed to influence a case’s outcome – mainly as a result of anchoring effects and the work-pressures faced by the other judges. Emphasis on group cohesiveness and collegiality – which is important for the functioning of the chamber – puts additional pressure on judges not to deviate far from the initial solution suggested by the rapporteur.
Section 4 conducts an empirical test of the effects of the legal origins of judges and panel dynamics on judgments.
The authors elaborate two hypotheses building on the analysis developed in the preceding sections. First, the legal origin of EU judges will influence their attitudes toward administrative appeals, with judges from French-influenced countries being more likely to be deferential to, and, therefore, rule in favour of the Commission than those from the rest of EU countries. The second hypothesis is that the rapporteur exerts a greater influence on case outcome than other judges do.
They test these hypotheses by examining whether there is a statistically significant effect between a judge’s legal origin and case outcomes, using the judge’s country of origin as a proxy for the legal tradition in which she was educated. Drawing from the CJEU’s case law database, the authors created a dataset comprising all judgments deciding appeals by private parties applying to annul a competition infringement decision (not including state-aid cases) by the Commission from the General Court’s inception on 1989 to September 1, 2015. They also created a dataset of the background characteristics of the judges who have served at the General Court, on the basis of publicly available information. The authors then control for a judge’s prior work experience, legal training in French-speaking countries and other control variables that one may expect to use in regression analysis, such as gender and age. In addition, the authors control a number of potential external factors that may influence case decisions, such as case complexity (using judgment length as a proxy), type of case, Member State intervention, time period and ideology (using the political orientation of the government that appointed the judge as a proxy).
They find that chambers with more strong French-influenced judges are more likely to rule in favour of the Commission. They also find that there is a noticeable difference in approaches between Nordic and common law, on the one hand, and other civil law systems (pure French, German and ex-socialist) – with Nordic and common law judges both being significantly likelier to annul infringement decisions than judges from other jurisdictions. Regarding judge rapporteurs, they find that the odds of a ruling being in favour of the Commission increases substantially if the rapporteur hails from a jurisdiction with strong French influence. A simple example will illustrate the effects. If the rapporteur is from a country with strong French influence, there is a 37% chance that the General Court will annul (or partially annul) the Commission’s decision. If a judge from a country with weak French serves as the rapporteur instead, the odds increase to an 81% chance of annulment (or partial annulment).
This article is in line with an approach adopted by many US studies of judicial decision-making (and one that I adopted in my own academic research) – i.e. it starts from the premise that the personal characteristics of decision-makers affect their decisions; and that decision-makers are, in turn, moulded by their training. Indeed, one could argue that legal education is a multi-year process of technical indoctrination).
The challenge with studying judicial behaviour is that we lack good tools to do it. One option is thus to ignore such behaviour and focus on pure legal analysis – which, while allowing for some types of analytical and normative analysis common in the legal literature, is an approach that is usually unable to describe the actual reasons underpinning individual decisions. Another approach is to rely on behavioural data, either qualitative (e.g. interviews with judges or counsel) or quantitative (e.g. statistical studies of case law by reference to some criteria that are thought to have explanatory value). This latter approach has led to the creation of a cottage industry among legal economists, but, to my mind, it often fails to take seriously the very real constraints that legal tests, procedures and frameworks exert on judicial behaviour.
I like how this study, while falling squarely within the latter approach, tries to address these limitations by paying at least lip service to the idea that legal tests and procedures matters. However, the paper itself focuses on a very narrow point – albeit one which seems to have explanatory value. I am unable to comment on the statistical correctness and reliability of the study, but its results seem plausible – even if, as the authors note, it is not in line with what occurs in other international courts. Indeed, they note that political bias seems to matter more than legal-origin at the European Court of Human Rights. However, they argue that those court’s decisions are inherently political, while competition decisions are mostly technical. That may be the case, but I think this is an assertion that needs to be tested somehow before we accept it – even if I have no idea how to do that.