EU court strikes down the language regime of EPSO’s recruitment process

On 24 September the General Court annulled three competition notices which require the candidates to choose English, French or German as their second language and as the language of communication with EPSO, the European Personnel Selection Office, responsible for recruiting staff to work for the European Union (Joined Cases T-124/13 and T-191/13). Italy and Spain requested the General Court to annul the relevant competition notices (published in December 2012 and January 2013), arguing that these are discriminatory and infringe the EU’s language regime established by Regulation No. 1 of 1958. The case was decided by the Eighth Chamber of the General Court, composed of three judges. The main author of the judgment is presumably the Greek judge, Dimitrios Gratsias, who is indicated as judge rapporteur.

The three competition notices challenged by Italy and Spain concern the recruitment of assistants in the field of biology, other natural sciences and engineering, of administrators in the field of security of buildings, and of assistants in the field of audit and finance. As regards the limitation of the language of correspondence between EPSO and the candidates to three languages, the Court found a violation of the EU’s language regime. Even if EU institutions are allowed to determine the detailed rules for the language regime in their internal rules, according to the Court the competition notices cannot in any event be regarded as internal rules. Thus, in relation to the language of correspondence with EPSO, the Court does not go beyond a formal argument in its reasoning. On the other hand, as regards the obligation for the candidates to choose English, French or German as a second language for the competition, the Court’s ruling is based on a substantive argument: it constitutes discrimination on grounds of language.

The decision relies heavily on a precedent case which arose from similar facts. In that case the Court of Justice (CJEU) reversed a judgment of the General Court (Case C-566/10 P of 27 November 2012). According to one commentator (see Albert Sánchez Graells’ blog post), however, the General Court “went too far and emptied the analytical framework created by the CJEU in 2012 of any meaning”. From the 2012 precedent it seems clear that according to the CJEU it is for the EU institutions, and not for the court, to weigh the different interests of service against each other, and in particular to decide if the interest of limiting the number of languages of the recruitment competition outweighs the objective of employing the most competent candidates (see para. 94 and 97 of the judgment). In that occasion the Grand Chamber of the CJEU ruled that the interest of the service may be a legitimate objective justifying limitations on the principles of non-discrimination and proportionality (para. 88). The CJEU also mentioned the opportunities for recruited officials of learning, within the institutions, the languages necessary in the interest of the service (para. 97), as a possible objective to be weighed against the limitation of the number of languages already in the recruitment stage. In the present case the General Court finds that the Commission did not provide sufficient proof that limiting the choice of the second language and of the communication with EPSO to English, French or German is justified by a legitimate interest of service. Sánchez Graells points out that by doing so, the General Court might have gone too far, because it substituted EPSO’s discretion with its own. But how far the EU institutions’ discretion may go?

This discretion is not unlimited. The CJEU clearly stated that the interest of the service must be “objectively justified and the required level of knowledge of languages must be proportionate to the genuine needs of the service” (para. 88 of Case C-566/10 P). The CJEU itself had already applied this test in a previous case, where it examined whether the level of knowledge required is one appropriate to the actual requirements of the service (Küster v. Parliament, Case 79/74). In that occasion the Court accepted the justification offered by the European Parliament and dismissed the application. In the present case, after applying the same test, the General Court finds an infringement of the EU’s language regime and annuls the challenged competition notices. So I do not share the criticism expressed by Sánchez Graells, and do not think that the General Court’s decision would be incompatible with the CJEU’s approach. Even so, whether the CJEU will share the General Court’s reasoning is still an open question. The decision may be appealed before the CJEU within two months, thus by 24 November. From the case-law, however, it seems clear that EPSO’s decision to limit the number of languages that may be used in the recruitment process is subject to scrutiny by the EU courts.

It is also very interesting to have a look at the General Court’s analysis and at the grounds for its decision. The challenged competition notices state that English, German or French are the languages that newly recruited employees need to know in order to be immediately operational and able to communicate efficiently in their daily work. The reason provided by EPSO in the notices is that these three languages are the most widely used ones both in internal and external communications and in the handling of files. Moreover, they are also the most widely chosen by candidates of competitions as a second language. The competition notices also specify that, in the interest of equal treatment of all candidates, everyone has to take the tests in his or her second language, to be chosen among these three, including those who are native speakers of one of them.

The General Court dismisses all these arguments questioning the relevance of the statistical data presented by the Commission in support of the aforementioned statements. For example, the Commission provided statistical evidence that English, French and German are the three languages to which almost all documents are translated by the Directorate-General for Translation. The Court replies that the presented data refer to the practice of the Commission only, not to the specific institutions involved in the challenged competitions. Moreover, the numbers presented by the Commission do not distinguish between documents translated for internal use and those translated for external use. Therefore, it is not possible to identify either the percentage of texts of internal origin or the distribution of those texts among the different services which could allow to evaluate if these languages are really the most widely used ones in the sectors involved in the challenged competitions (para. 118-121).

Until this point the reasoning shows a very formalist approach and gives the impression that the General Court refuses to accept the obvious, requiring hard evidence of well-known facts. But the Court does not stop here. It actually carries out a detailed analysis of the statistical data presented by the Commission, and draws interesting conclusions from them. To make a long story short, the Court points out that from the statistics it emerges clearly that French and German are behind English with a huge gap, while Italian and Spanish are in the fourth and fifth positions respectively with only a very small gap. Therefore, these statistics do not justify the limitation of the required languages to English, French and German and the exclusion of others. They could, according to the Court, justify the requirement of the knowledge of English, but not that of French and German, which represent only a small percentage of the translated texts compared to English and are not much ahead other languages (para. 126).

The Commission also produced statistics concerning the main language of its officials and agents. It is not entirely clear what the Commission wanted to achieve by showing these data. The main language of the officials is their mother tongue, thus it reflects their nationality. The presentation of these data is rather counter-productive, since they reveal the disproportionate presence of French-, German-, Dutch- and English-speaking officials and agents in the EU institutions (in this order). They represent a total of 56.3%. Moreover, Italian-speaking officials represent roughly the same percentage (9%) as officials whose main language is English, Dutch or German. French-speaking officials (presumably including French, Belgian and Luxembourgish nationals) represent 26.9% (para. 129-130). So in this respect English is not the dominant language. Other statistics prove that while 56.4% of the officials and agents have English and 19.8% of them French as their second language, this percentage is only 5.5% for German. These data, however, the Court argues, have limited significance, as they do not consider the possible third or fourth languages spoken by the officials (para. 132-133). Consequently, these percentages are probably even higher. In any event, none of the statistics presented by the Commission seem to justify the requirement of the knowledge of German, the Court says. If the knowledge of German can be required, then also requiring the knowledge of Italian, Spanish or Dutch is justified (para. 134).

The General Court concludes that limiting the choice of the second language by candidates to English, French or German is neither justified nor proportionate to the goal of recruiting immediately operational officials and agents (para. 145). However, it seems that the Court could accept the limitation to the sole English language, which raises some interesting questions. If the CJEU will uphold the General Court’s decision, it will mean that the European Union cannot pick a few languages as “preferred languages” from the bunch of 24 official languages. Either it may move into the direction of choosing one common language (in all likelihood English), or it has to treat all the official languages of the Union equally, with all the costs this entails.

End note: Ironically enough, the text of the judgment is available in several official languages of the EU but not in English or German. This comment has been written on the basis of the Italian language version.