Defining a set of conduct as a single and continuous infringement (SCI) has significant consequences for parties to a cartel, in particular for their exposure in follow-on damages actions. It is no surprise, therefore, that the increased prevalence of damages actions in Europe has led to greater scrutiny of the Commission’s findings about SCI.

In several recent judgements the General Court and Court of Justice have grappled with what to do about cases where the Commission has found a particular company to have participated in only part of a broader set of infringements, that broader set constituting – in the Commission’s view – a single and continuous infringement. We may, for example, be looking at a case where a company participated in one product of a two-or-more product cartel, or where it was a fringe player in terms of either its participation, or its geographic reach.

Consequences of a Finding of SCI

As a reminder, the finding of a single and continuous infringement has broadly three potential consequences for a case.

First, prescription: if earlier conduct can be regarded as a single infringement with later conduct, then earlier conduct that might otherwise be prescribed would still be attackable.

Second, fines: if there is one single infringement, then there is one fine, one entry fee and one ten percent cap. If there are two infringements, then there can be two fines, with two entry fees, each subject to the ten percent cap.

Third, damages: participants in a single and continuous infringement are jointly and severally liable for the damage that they have caused.

If this is a classic multilateral around the table cartel, then the characterisation of the SCI and each undertaking’s part in it is relatively straightforward. The difficulties arise when the participation – and liability – is asymmetric.

So a simplified example would be two undertakings who are deeply involved in all aspects of a cartel, compared to a third which is only peripherally involved, whether that peripheral nature is due to geography, product scope or conduct or potentially other factors.

In the past, when public enforcement was essentially the only concern of defendants, then provided that the fine was appropriately tailored to the particular undertaking’s participation, the finding of that undertaking’s participation in an SCI was of little importance to that undertaking.

In an era of damages actions obviously that changes significantly. A small, peripheral player found to have been part of an SCI in a major cartel is potentially on the hook for all of the damages caused by the cartel as a result of its joint and several liability for the SCI. In practice this is very unlikely to happen (as among other things it would likely require the bankruptcy of the other players), but having to pay a greater share of the overall damages than it believes appropriate is a much more likely scenario.

Because of this concern over the implications for private enforcement, we now have a set of cases – Coppens, Aalberts and Soliver – where the Court expressed discontent, in various ways, as to how the Commission characterised the participation of certain undertakings in the single and continuous infringement.

In Coppens, the Court of Justice found that a company had participated in one product of a two product cartel, and was not a part of the two-product single and continuous infringement. Nevertheless a finding of infringement in respect of that one product was upheld.

In Aalberts, the Court took a different approach, finding that a company had only participated in some of the anti-competitive contacts. The Court – relying on the Commission’s characterisation of the contacts as a single indivisible infringement – annulled the entire decision.

Then in Soliver, the Court found that a fringe player carried out acts that fell short of full complicity in the more extensive single infringement carried out by other players, recognised that those acts were clearly anti-competitive, but nevertheless annulled the decision as a whole indicating that although the anti-competitive nature of those acts was clear, the party had not had an opportunity to respond to objections focused explicitly on those anticompetitive acts during the administrative procedure.

These cases taken together raise two distinct issues in terms of Commission decisions.

First, how should the Commission properly characterise the role of fringe players, or players who only participate in some aspects of the anti-competitive conduct. i.e. If we take the facts of, say, Soliver, how should the Commission have written its decisions in order to meet the approval of the Court. And it’s important to bear in mind here that it’s not (necessarily) just a question of how the Soliver decision should have been written, but also how the other decisions related to that cartel – that affecting St Gobain for example – should have been written.

Second, and allowing for the fact that these are complex issues of fact where reasonable people can disagree, should the Commission improve the way it drafts its decisions such that, if the Court disagrees on the particular characterisation of the single and continuous infringement, the entire decision does not necessarily fall but would be upheld in part.

Properly Characterising the Role of Fringe Players

Looking at the first of these, how should the Commission properly characterise the role of fringe players, together with those that are involved in conduct with fringe players. I do not think Soliver is particularly clear on how the subsequent Commission decision should be drafted.

Let’s take the example of two companies who are central to the arrangements, and one peripheral player.

Soliver says that we should not too quickly find that the peripheral player was liable for the single and continuous infringement. But what does that mean for how the infringement should have been drafted in the first place?

We know that for the peripheral player, they should not be described as being part of the SCI. But what should they be described as? Participating in an infringement separate to the SCI is obviously one possibility. But that has implications for the main players and the drafting of their decision? Are they now implicated in two separate infringements? The main one – the previously characterised SCI less that aspect where the fringe player participated – and then a second one with the peripheral player?

If this is right, then in the Soliver case, should St Gobain have been characterised as being in two infringements? This seems to be simple enough and may work in some cases. But what if, from the point of view of St Gobain, the conduct all looks very closely connected? From its point of view it looks like a single infringement? If from St Gobain’s perspective all of the essentially similar conduct covered the same territory, the same product, with the same anti-competitive object, then is it appropriate to regard that as two infringements simply because it also involved a peripheral player?

And here we have potentially the reverse problem of the peripheral player.

As I said, for the peripheral player, the implications of the finding of the SCI may be limited for public enforcement, assuming the fine was appropriately tailored, but the implications for follow on damages may be significant.

For the main player, there may be no real consequence for follow on damages – whether they are sued for damages for two infringements or one may make little difference as the cumulative damage must be the same. But there may be real consequences for public enforcement. For example if the main player was facing a capped fine for the single infringement, is it really open to the Commission to impose two fines, each with their own 10% entry fee and 10% cap simply because there was a peripheral player involved in part of the conduct?

This concern is perhaps clearly evident when looking at an example of an infringement covering multiple Member States. If these were a single infringement, then there would be a single fine; if there were multiple infringements, then there would be multiple fines with multiple 10% caps. If these were long running infringements, then the total fine could be well in excess of 10% of turnover.

That cannot be the correct result. If two multi-national companies are involved in a cartel covering five Member States, and in each of those Member States they cartelised together with a local company, then that should not lead to a finding of five separate infringements, with five fines and five applications of the 10% cap.

So defining multiple infringements may work for the peripheral player, but the main player risks coming out of this rather badly.

So the right answer is not, I think, to simply define multiple infringements.

Rather it is to define one single infringement – assuming the evidence points in that direction for at least one of the players – and then for those companies that are not involved in the full infringement, to define their participation as being of “part of the single infringement” whether that part is in relation to particular products, particular conduct, or particular territories.

So we might have, for example:
Company A liable for an SCI encompassing ten Member States and three products;
Company B liable for an SCI encompassing ten Member States and only one product;
Company C liable for an SCI encompassing five Member States and three products;
And so on.

Now there are two very different judges that have to be satisfied with this formulation:

• the judge that reviews the public enforcement decision and needs to see if the liability of all players is properly analysed;
• the judge that is charged with determining damages on the basis of a follow on action where the judge is bound by a finding of joint and several liability in respect of “the infringement”.

For the first, the General Court and Court of Justice have accepted this type of reasoning in the Bananas cases. But for the second, it is an open question – and likely will be for some years – as to whether this formulation would be sufficient for a national judge to delimit the joint and several liability of a cartelist to that “part” of the infringement which the Commission has defined in the decision.

I hope it would be for two reasons.

First, it makes sense to me because it is analogous to participants being involved for different duration, and a judge would deal with that easily enough with no conceptual difficulty of an undertaking being involved in – for example – only the last five years of a ten year cartel. Surely a judge looking at a Commission decision that, for example, finds an undertaking to have participated in a pan-European cartel save for Germany, would find no real conceptual difficulty in excluding that company from damage caused in Germany.

The second reason I hope it would be acceptable is that I don’t see a better alternative.

Drafting a Decision: Protecting Against Total Annulment

So much for how the Commission should draft its decisions to reflect properly each cartelist’s liability.

Now I want to turn to the question of whether the Commission should take greater steps to draft its decisions in such a way that, if a court disagreed with its characterisation of the single infringement, the Commission would not be facing total annulment but only partial.

Now of course making Commission decisions more resistant to Court review is not historically something that the defendant bar is interested in, but here again I think private enforcement is making a difference to incentives.

For example, imagine once again there are three undertakings implicated in a cartel. Two are clearly central players, one is implicated in only part of the infringement (be it the conduct, the product or the geographic area). The Commission decision characterises all as participants in the single and continuous infringement, the Court disagrees and annuls the decision against the third player in its entirety. Essentially what happened in Soliver.

If we assume that the Commission decision will be followed by follow on damages actions, then the defending lawyers of the main players have two objectives. First, to annul the Commission decision as regards their own client. But second, if the annulment action by the main players fail, they want to make sure that the Commission decision against the peripheral player survives – or at least as much of it as possible does.

So you do not want the decision against that defendant to be struck down because the Commission inappropriately characterised the infringement and the liability, and the Commission drafted the decision so that the entire decision falls if part of it does.

When the Court annulled Soliver it did not do so on the basis that Soliver’s conduct was in fact not anti-competitive. Quite the opposite. The Court clearly indicated that that part of the overall conduct in which Soliver was involved was itself anti-competitive. However the Commission had not characterised that part of the conduct as being anti-competitive in the Statement of Objections or the Decision. So Soliver had not been able to exercise its rights of defence in relation to that part of the conduct by itself.

This contrasts with the reasoning used in Aalberts, where the Court relied on the Commission’s own characterisation of the single conduct as a reason to annul the decision in its entirety. In that case, the Court again disagreeing that Aalberts had been involved in a single infringement encompassing all of the anti-competitive conduct, the Court quoted standard text used by the Commission against the decision itself – if it would, as the decision stated, be artificial to split the conduct into separate infringements, then if part falls, it all falls.

Soliver takes this substantive law argument, and turns it into – or possibly adds to it – a procedural one. What Soliver tells us is that unless the defendant has been adequately able to defend itself against the allegation that its conduct – being something less than the full SCI – was nevertheless anti-competitive then the decision will be struck down in its entirety.

So if the Commission defines a single and continuous infringement, it should also address and characterise “parts” of the single and continuous infringement conduct that could be regarded as standalone infringements.

This might get complicated.

What if the conduct covered ten member states, and the Commission considered it to be a single infringement covering all ten? What level of reasoning in the decision is necessary to justify the possibility that each Member State constituted a separate infringement? A full analysis for each individual Member State on a standalone basis?What about combinations? Benelux might be one infringement, Iberian peninsula might be another. How many alternatives do we need to envisage?

Or what if three products were involved, and again the Commission found that this was a single infringement, but the Court disagreed? Three full sets of legal analysis for the alternative that each product was a separate infringement?

What if we combine the two?
Do we then need to do an analysis of the restrictive effects of the conduct in respect of each of the three products in each of the ten Member States? That’s thirty separate infringements even before we get into the question of regional rather than national infringements.

It doesn’t take much imagination to see that even spotting each possible variant is an onerous task. And full analysis of each possible variant would increase the length of Statements of Objections and decisions quite considerably.

It seems unlikely that the Court had this in mind.

I suspect a more efficient approach, while still acceptable, is to make a judgment call on what needs to be done in each case.

If there is substantially the same conduct and market situation in each Member State, for example, then it’s probably acceptable to indicate rather briefly that although the Commission considers the single infringement to cover all ten Member States, conduct in any one member State, or combination of Member States, would similarly be an infringement.

It’s possible to imagine more complex situations, however, that might need greater elaboration.

An argument in favour of this moderate approach in most cases – aside from the sanity of all those involved – is that the very concept of a single infringement necessarily implies the existence of a set of instances of conduct that analysed separately could constitute separate infringements. In most cases, a simple mention of this should be sufficient to put a defendant on notice that in exercising their rights of defence they need to reply to that possibility.

(This presentation was given at events at Lincoln’s Inn in London and Brussels Matters in Brussels in late 2015. I am grateful to Thomas Sharpe QC, James Flynn QC, and Conor Maguire for the invitations to speak.)