Case law Developments in PIL Views and comments

EU Private International Law before the ECJ: Back to Procedural Issues

This is the last post in the series dedicated to the empirical analysis of the ECJ’s case law in the field of EUPIL. The previous posts can be found here and here.

This post is slightly different from its predecessors, as the angle of analysis is reversed. Rather than (just) analysing the characteristics of the ECJ’s case law in the field of EUPIL, this post purports to use such case law as an indicator of the transformations in the working methods of the ECJ itself. I refer back to my previous posts as concerns the methodology and definitions upon which this research is based.

The starting point of my analysis is the objective set out by Recital 6 of the Court’s Rules of Procedure (RoP): these aim at “maintain[ing] the Court’s capacity, in the face of an ever-increasing caseload, to dispose within a reasonable period of time of the cases brought before it”.

Chart 1  below shows, in this respect, that this objective is being pursued by the Court in a rather effective manner.

Chart 1

The red line in Chart 1 indicates the evolution over time of the Court of Justice’s overall workload (and not only of preliminary reference procedures). The numbers on the vertical axis shall therefore be interpreted as indicating the total amount of cases filed each year. As I could not find any official statistics pertaining to the 70s, 80s and early 90s, this data was obtained very pragmatically. I used the “advanced search” form on for each year since 1976, I selected the time frame 01/01 to 31/12, filtering the results based on the type of court (Court of Justice), and the type of date (date of the lodging of the application initiating the proceedings). The red line portrays the results thus obtained. For the sake of consistency, I used this methodology for all the years between 1976 and 2022, even if official statistics are available since 1997. The divergence between the two sets of data (official and unofficial) is negligible (< 5 per year).

What happened in 1979? I am actually not sure. It looks like a huge number of cases on the status, remuneration and benefits of officials were filed that year. For most of these cases, there is no judgment, which probably means they were withdrawn at some point. Their effective impact on the workload of the ECJ remains therefore undetermined.

The line in dark green shows the average length (expressed in calendar days) of preliminary reference procedures in the field of EUPIL (this data is global, as it refers to cases decided with and without an Opinion of the AG, as well as cases that have been withdrawn and removed from the register).

The line in lighter green (which overlaps with the former until 2001) portrays the average length of preliminary references decided with an Opinion of the AG, the line in blue those which dispensed with it. The interruptions in the latter mean that there were no cases decided without an AG’s Opinion in the corresponding years.

Finally, the line in violet represents the average length of urgent preliminary reference procedures (PPUs) in the field of EUPIL. These cases, all dealing with family law, are decided with the support of the AG Opinion (formerly, a View) and a hearing. The average length of the proceedings remains remarkably low: to the present days, 80 calendar days (on this topic, see also this document).

Against this backdrop, the objective set out by Recital 6 seems met: the average length of (ordinary) preliminary reference procedures has been following, over the last years, a decreasing trend. How could the Court manage such result, despite its increasing workload?

Of course, there have been important institutional changes over these four decades: the progressive enlargements of the EU, the devolution of certain competences to the General Court and to the Civil Servant Tribunal (and back again) have all had an indisputable impact on the Court of Justice’s caseload. The purpose of this post, however, is to demonstrate that much was done also from the standpoint of internal (re)organization and working methods. In this respect, the analysis of the procedural treatment reserved over time to EUPIL preliminary references shows the noteworthy adaptability of the Court of Justice’s internal functioning and its ability to optimize the use of its resources. As we will see, there have been significant transformations as concerns the use of judicial formations (A), of the AGs’ Opinions (B) and of hearings (C). This will also be the opportunity to come back on the issue of informal specialization of the Members of the Court, which I remarked in my first post (D).

 A. The Transformations of Judicial Formations

Two important observations can be drawn from the stock chart below: first, EUPIL preliminary references have always represented a negligible part of the ECJ’s total caseload, having amounted to less than ten cases per year until the early 2000s. Second, there has been a considerable shift, over the years, as concerns the judicial formation adopted by the ECJ to decide on the questions raised by these cases.

Chart 2

In the early days of the ECJ’s activity under the 1971 Protocol on the interpretation of the 1968 Brussels Convention, most of the EUPIL preliminary references were decided by the full court. It must be assumed this was a clear and conscious stance taken by that Court with respect to EUPIL cases, and not just the indirect result of a different era, when the ECJ, counted only nine Members and had a very limited caseload, thus having the opportunity to resort to the plenum as the default judicial formation. To the contrary, it is apparent from the judgment rendered in Tessili that the Court could already operate in smaller deciding panels (two “Presidents of Chambers” are mentioned in the part of the decision listing the composition of the court).

The preference for the Full Court, manifested by this early case law, should come as no surprise: the cases decided by this formation between 1976 and 1980 (De Bloos, Mines des Potasse, LTU, to name a few) laid the foundations of modern EUPIL, defining extremely important methodological and terminological issues that still shape today’s way of approaching the new generation of EUPIL Regulations.

What was that “Full Court”, however? It was certainly nothing similar to today’s Full Court, regulated by Article 60 RoP and Article 16 of the Statute. It was admittedly surprising to note that the Full Court of the early days consisted of sometimes 9, sometimes 7 judges, following patterns whose underpinning logic is not immediately perceivable by the external observer. It looks like this “Full Court” was indeed a rather flexible judicial formation, counting a “bigger” and a “smaller” plenum, corresponding in essence to what we call today Full Court and Grand Chamber (I am drawing this information by this scholarly article of 2001).

With the exception of the initial period going from mid-70s until the 80s and another intermission in the early 90s, Chambers of five judges have remained the most common judicial formation for EUPIL cases. The first EUPIL preliminary reference deferred a Chamber of three judges was case 120/79, on maintenance obligations. Since then, this judicial formation has been seldom employed throughout three decades, having become more recurrent over the last years. This can be seen as an integral part of the ECJ’s overall attempt to optimize the use of its resources, including its personnel. Only 6 (7,6 %) of the cases deferred to a Chamber of three since 2003 was decided with the support of the Opinion (2003 being the point in time when the AG’s Opinion was no longer systematically required for all cases: see infra Section B). In practice, this means that these cases did not raise legal questions that, owing to their novelty, importance or technical complexity, called for the advisory intervention of the AG. A Chamber of three is overall more efficient when deciding this type of cases, insofar as the average length of the proceedings before it is 337 calendar days, compared to 437 calendar days that are needed, on average, by a Chamber of five to adjudicate without an Opinion.

A final word on the Grand Chamber which, as we know it, was created in 2003. Owusu was the first EUPIL preliminary reference assigned to this judicial formation, which has been used rather sparingly over time (only 3.8 % of EUPIL preliminary references were assigned to it). The period between 2006 and 2009 was marked, however, by a veritable boom of Grand Chamber cases. This was, after all, an “era of first times”: the first ever preliminary references on the Brussels IIbis Regulation (C-405/06) and on the 1980 Rome Convention (cases ICF and Koelzsch), as well as the first occasion for the ECJ to test the Brussels regime against the challenges brought along by the Internet (cases Pammer and Alpenhof, eDate).

B. The Opinion of the AG

Speaking of eDate, have you ever noticed that its “ancestor”, Shevill, has not one, but two Opinions, delivered by two different AGs? Same things for Marinari, also filed in 1993. As correctly indicated by AG Léger, it could “infrequently happe[n]…, by reason of the reopening of the oral procedure and as a result of happenstance in the order of business of the Court”, that two Opinions are delivered in the same case. The Shevill judgment explains, in this respect, that the case was initially assigned to the Sixth Chamber of the Court (chamber of five) and referred, after hearing the Opinion of AG Darmon in July 1994, « back to the Court », meaning the Full Court. The oral phase of the procedure was consequently reopened before this bigger judicial formation, and a new Opinion was delivered by a different AG, Mr. Léger. In Marinari, the issue was, again, the reopening of the oral phase of the procedure, without any referral to a different judicial formation. Again, two different AGs delivered an Opinion in the case. The fact triggering the second intervention of the AG is, therefore, the reopening of the oral phase of the procedure as such, and not the referral to a different judicial formation.

While the merits of having two different AGs delivering an Opinion in the same case could lie in the potentially different point of view introduced into the debate, thanks to a “fresh start” to the study of the case file, this working method could be deemed inefficient insofar as at least four different persons (the two AGs and their respective référendaires) are called to work on the same case from scratch (in practice, the two AGs adopted the same stance in both Marinari and Shevill).

The reopening of the oral procedure is only ordered in exceptional circumstances and is not a common occurrence. This has not happened again in EUPIL cases since 1993, but it could potentially happen. The new RoP provide, in Article 83, that the Court may at any time order the opening or reopening of the oral part of the procedure, in particular if it considers that it lacks sufficient information or where a party has submitted a new fact which is of such a nature as to be a decisive factor for the decision of the Court, or where the case must be decided on the basis of an argument which has not been debated between the parties or the interested persons referred to in Article 23 of the Statute. It is worth noting, however, that today the reopening of the oral phase of the procedure no longer entails the intervention of two different Advocates General. I can mention two cases, both outside the field of EUPIL, where an order under Article 83 RoP was adopted: C-168/16 and, more recently, C-530/20. In both, a single AG delivered two subsequent Opinions. It seems therefore that the Court is nowadays favouring efficiency over the plurality of views, consistently with the general objective of reducing the length of the proceedings set out in Recital 6 RoP. Also noteworthy is that the involvement of a single AG in each case is now provided also for the delivery of Opinions (in French, Avis) requested in accordance with Article 218 (11) TFEU (Recital 5 RoP).

The biggest innovation concerning the role of the AG –  also made in the attempt to increasing the ECJ’s overall efficiency – happened in 2003. Before this date, the AG had to deliver an Opinion for all preliminary references brought before the Court. This explains why, up to that moment, 100% of the EUPIL preliminary references decided by a chamber of three judges came with an Opinion, whereas only the 7,6 % of the cases assigned to such judicial formation after 2003 called for the AG’s advisory intervention.

Nonetheless, Chart 3 below demonstrates that the great majority of EUPIL preliminary references is decided, even after 2003, with the support of the AG’s Opinion.

Chart 3

Of all EUPIL cases having dispensed with an Opinion, 60 % have been assigned to a chamber of three, and 40 % to a chamber of five. 16% have been decided through a reasoned order under Article 99 RoP (all of them adopted by a Chamber of three, except for C-518/99). The possibility to define a case by means of a reasoned order explains the existence of a certain number of cases decided without an Opinion even before 2003.

C. Hearings

Another area where the Court has striven to increase its efficiency concerns the holding of hearings. According to Recital 6 to the RoP, “in order to maintain the Court’s capacity, in the face of an ever-increasing caseload, to dispose within a reasonable period of time of the cases brought before it, it is also necessary to continue the efforts made to reduce the duration of proceedings before the Court, in particular by … providing for the Court to be able to rule without a hearing if it considers that it has sufficient information on the basis of all the written observations lodged in a case”. As I mentioned in my previous post, a hearing shall be held, according to Article 76 RoP, when it has been requested by an interested person that has not participated in the written phase of the procedure.

Chart 4 below shows the evolution in the use of hearings in EUPIL preliminary reference procedures.

Chart 4

The analysis of more than forty years of case law in a given field of law is also a journey through different drafting styles, used by the Court in its judgments. This is why, in a certain number of cases, it was not possible to determine whether or not a hearing was held. This concerns, in particular, the cases filed between 1984 and 1985. More recently, a certain number of judgments only mention the observations of the parties, without referring either to a “written procedure” or, more explicitly, to “a hearing”. Where there was no AG Opinion, or when this did not clarify this point, these cases were also classified in the “unsettled” category.

This said, it must be noted that the recent trend goes, quite indisputably, towards reducing the number of hearings held in EUPIL cases. Intuitively, holding a hearing will delay the procedure, and it makes sense to limit this effect to the cases where an oral procedure is necessary for the correct understanding of either the legal questions referred to the Court or of the context in which they were raised, as well as in the cases where it serves to preserve the right to be heard of the parties and the interested persons listed in Art. 23 of the Statute. Overall, hearings have been held in 14% of the cases assigned to a Chamber of three and in 62% of the cases decided by a Chamber of five. This percentage drops to 54.6 % in cases decided by a Chamber of five after the current RoP have come into force. There is, however, a certain number of EUPIL cases decided before 2012, whose judgment only contains references to the written phase of the procedure. It must be assumed that, therein, a hearing was not held, and that the possibility to dispense with the oral procedure existed also under the previous RoP.

D. The Specialization of Judges and AGs

To conclude this survey of the transformations made, in the quest for more efficiency, to the working methods of the ECJ, I wish to come back to the issue of informal specialization of judges and AGs, which I remarked in my first post, focusing solely on the 2015-2022 time frame.

I came back to this issue with ambivalent feelings, and I do not have any conclusive opinion on this topic, although I am keen on confirming my initial impression. Chart 5, below, shows the rate of intervention of different AGs in EUPIL cases since 1976.

Chart 5

The picture is indeed quite fragmented, but two observations are in order. First, just eight AGs have been in charge of 50% of the total EUPIL cases (right side of the pie chart), whereas the other 50% of cases is shared between 51 different AGs. Second, the eight AGs on the right side of the chart have all exercised their functions in recent times (late AG Bot, who was the first among them to arrive at the Court, was appointed in 2006). It could therefore be concluded that specialization of AGs – if any – is a relatively recent trend, with the last 15 years testifying of a certain tendency to see a smaller number of AGs systematically involved in EUPIL cases.

Chart 6 below is a variation of Chart 5, taking into account the evolution over time of appointments of AGs to EUPIL preliminary references (click here for a slightly larger picture).

Chart 6

In the attempt to increase the readability of the chart, only AGs having been appointed in more than five EUPIL cases have been named. The category “others”, in yellow, accounts for the remaining cases and groups 27 different AGs (for 74 cases). As remarked above, the specialization appears stronger in recent times, with the yellow category disappearing completely between 2011 and 2017. The recent spike in the yellow category has a clear explanation. AGs Jääskinen, Saugmandsgaard Øe and Bobek, who have been highly active in the field of EUPIL, have ceased their functions in 2019 and 2021 respectively. We are now, it seems, in a phase of transition, where new AGs have taken over and might develop, in the coming months/years, a similar informal specialization in EUPIL cases. Quite remarkable is, in this respect, AG Pikamäe, who already appears in the Chart despite his recent appointment.

The exact same situation exists with respect to Reporting Judges, with the notable difference that only two of the judges appearing on the right side of the pie chart are presently still working at the Court. In this domain domain, the turnover effect will be even higher in the coming months.

Chart 7

Chart 8. Click here for a slightly larger picture.


As I already mentioned, the specialization of AGs and Reporting Judges, if any, is purely informal, and should be taken as an objective data emerging from the analysis of existing case law: some among them have simply dealt with EUPIL cases more often than others. This approach could favour internal efficiency, since prior dealings with a certain subject matter could reduce the time needed for assessing the case and take a stance on the legal question it raises. It remains, at the same time, flexible enough to ensure the correct functioning of the Court (for example in terms of equitable distribution of cases among judges/AGs and the prompt dealing of PPUs and PPAs). A more rigid approach to specialization (such as the formal institution of specialized chambers) might jeopardize the achievement of this second “organizational” objective.

Legal Assistant at the ECJ. The opinions expressed in this publication are those of the author. They do not purport to reflect the opinions or views of the ECJ or its Members.

%d bloggers like this: