Big data and competition – big deal or no deal?

Jay Modrall: Richard one of the hottest topics in antitrust circles today is the role of big data and algorithms but a senior commission official said to me the other day that there are more conferences than there are cases. How do you see this area evolving?

Richard Whish: Well, he was certainly correct, there seems to me to be an incredible number of conferences, seminars, etc on this. At the moment every conference organiser has had, or is going to hold one. We at Kings College in February, had a major event for an entire day with speakers from all over the world and 200 plus people in the audience. So it is certainly extremely interesting and there haven’t yet been many cases. My view, in relation to this whole topic is basically keep calm. What we need to do is to sit down and work out where there are issues, are some of these issues ephemeral or not really as difficult as people like to make them out to be. Then we can move on to the next phase.

Jay Modrall: If and when they do, what areas and what practices do you think the antitrust authorities will go after?

Richard Whish: Well, I think there are some practices that involve the use of algorithms, price-tracking software, etc, which I think can quite clearly be unlawful by reference to existing competition law principles. A price-fixing cartel has always needed machinery to make sure that everyone is complying with the terms of the deal, monitoring what’s going on, that might have been done 50 years ago by a trade association. Today, it might be done by price-tracking software – I don’t actually see any difference between the two and if I try to impose a resale price maintenance on my dealers, I want to make sure that they are not cheating on the arrangement, and again I might use price-tracking software to carry out that policing function, but I don’t really see that there is anything new in a case like that.

Jay Modrall: Your reference to software brings up another question. People often refer to big data and algorithms in the same breath, almost interchangeably, do you see differences in the issues that these two areas raise?

Richard Whish: Well, yes I do and I think we’ve got to unpack the different issues and decide what problems can algorithms give rise to, what problems arise from big data. As far as the algorithms are concerned, one of the things that excites people is whether if I’m the user of software and I use algorithms and suppose those algorithms learn through artificial intelligence how to reach a collusive outcome in a market place, then am I, the user of the software, liable? Well, I think the answer to that is quite simple – yes, as a user of software I am responsible for what it does in the same way that an employer is responsible for what his or her employee does. But that doesn’t help us to determine whether there’s an illegal agreement or concerted practice because if my software simply learns how to track my competitors prices and then to raise them, is that not me, through my software, intelligently adapting to the behaviour of my competitor? The Court of Justice in Dyestuffs in 1973 said that a competitor is entitled to intelligently adapt itself to competitors and the Court repeated that in Woodpulp, so I don’t actually see that there is anything particularly problematic in that situation.

Jay Modrall: Well, while we’re waiting for the antitrust authorities to catch up, some regulators are racing ahead. The Commission, as you know, has proposed mandating data sharing, data access in the internetof things.  What do you see as the pros and cons of regulatory responses as compared to antitrust enforcement?

Richard Whish: Well I mean, one thing in my career in competition law has taught me is, that competition law cannot provide an answer to all the problems of the world. And it seems to me that when we come to the issue of super platforms and the vast data that they hold and the privacy issues that there might be and so on and so forth, I think some of these problems are, frankly, beyond competition law. And I understand in that sense why other legislative or regulatory responses might be needed. I’m not assuming that there’s need for legislation or regulation but there could very well be. On the other hand, the principles of competition law, and in particular the principles behind Article 102, are extensive and I can see that certain practices could be abusive in the sense of existing law and policy.

Jay Modrall: Well, we could talk for hours about the interplay between privacy and competition law alone, but we have to wrap it up so I’m going to ask you the inevitable Brexit question. How do you see Brexit impacting the evolution of the law in this area, or do you see the Commission becoming more interventionist, the CMA being more permissive?

Richard Whish: Well, I think there’s little doubt that the UK has had a major influence on the direction of competition law and policy in the EU over the last 20 years, maybe longer. I think we have encouraged people to adopt a market based consumer welfare approach towards the use of the law and not every other member State has necessarily enjoyed that terribly much. So, I think that when we are outside the EU, clearly our policy might develop in a different direction. We don’t know yet, but it might. I would say, however, that even if the CMA had an appetite for delving into the world of artificial intelligence and big data and so on, it has very little knowledge of these sectors whatsoever, because in recent years, all the cases involving Google and Apple and Amazon, etc, have been conducted in Brussels. So there’s a big learning curve, it seems to me, for the CMA in the future, As for the 27 member States and the Commission, who’s to say what will happen to competition policy but I could imagine, how it might become more interventionist, more quasi or even actually regulatory. I think that must be a possibility.

Jay Modrall: Who’s to say indeed. Thank you very much Richard.

Richard Whish: Thank you Jay.