As I self-advertised in my previous post, I participated yesterday at the European Data Protection Supervisor’s impressive workshop on Privacy, Consumers, Competition and Big Data, where, by the way, this blog received a few mentions.
My impression is that it provided a useful opportunity for various actors to reflect together on the nature, potential and limitations of each discipline in the wake of the EDPS preliminary opinion on these issues.
The workshop touched on competition issues several times. On the EU side, Kris Dekeyser gave the Commission’s view, and on the private side I was honored (and, frankly, a bit surprised) to be the sole EU competition lawyer speaking.
Julie Brill (FTC Commissioner; her speech is available here) and Pamela Jones Harbour (former FTC Commissioner now in private practice) also shared their views on the US approach to these issues.
I was asked to explain to a non-expert audience (by non-experts I mean those who retain the ability to realize sometimes that the king may sometimes be naked…) the notion of market power, why it is important for the application of our rules, how it is assessed in practice, and what are the particular challenges posed by digital markets and big data in this regard.
I’ll spare you the content on my intervention on the most basic issues; suffice it to say that I pointed out that the traditional means to define markets and market power are far from perfect in many ways, but that they’re not supposed to be used mechanically and in the abstract, that the Commission may depart from standard assessment tools to capture the dynamics of competition in any given sector, and that it enjoys wide discretion to act flexibly in this regard.
Moving on to the more interesting stuff. Following a conventional explanation of the main peculiar features of technology/digital markets and of their mixed competition law implications I gave my (non data protection expert) views on the big relevant issues addressed in the workshop, namely (A) What are the implications of data and big data for market definition and market power assessments and (B) Should privacy data protection standards be incorporated to substantive assessment under the competition rules
Today we’ll discuss A, and
tomorrow [on Friday] we’ll deal with B, so:
What are the implications of data and big data for market definition and market power assessments?
(i) Data is without doubt an increasing important asset/input, and it should no doubts be acknowledged as such. As some of you may remember, some time ago I commented on an article that essentially posited this idea, which I consider to be fairly uncontroversial. In this sense, I’ve no objection to the idea that, depending on the circumstances, data-related issues may give rise to competition concerns.
At the same time, however, data is an important asset or even crucial asset, but no more; and I don’t see why competition law would be required to adapt its rules to when applying them to data-heavy markets.
(ii) I see one exception to the above. As I explained in a recent post, our current turnover thresholds are not well-suited to capture mergers in the subsidized side of two-sided markets (which may often be markets where non-traded data is important). Only jurisdictions envisaging market share thresholds (often criticized, also by me) may be competent to assess these transactions. Facebook’s very recent decision to try to have the EU review the acquisition of Whatsapp is to be read within this context. I don’t know what the solution is, but it’s worth a thought.
(iii) Some (including Pamela Jones Harbour in her dissent to the FTC’s Google/Double Click decision) have advocated for a definition of relevant markets for “data used for x [in that case targetted advertising] purposes”. I’m not persuaded by this proposal (except perhaps when the data is subject to trade) because I’m not sure the intermediate data market is a meaningful market in the sense of competition law. If the alleged problem is that the use of data might have consequences in some markets, then my take is that it makes more sense to assess those markets directly.
(iv) Regarding the big substantive issue, which is related to scale, aggregation, network effects playing to the benefit of allegedly dominant firms, I essentially said that:
- far from being an obvious competitive problem this also has mixed implications, for data can also be a source of very significant efficiencies (and big data a source of big efficiencies) in many and important fronts;
- it is true that access to data may in some circumstances be a barrier to entry and even a very important one depending on the facts (I also noted that barriers to entry are not in themselves a problem requiring intervention because competition law is about conducts and not structure);
- many people throw out “essential facility” as a buzzword in this context to support the contention that some firms should be mandated to share data. In my view the term is used too loosely. As I explained, the identification of an essential facility is subject to an extremely high legal burden (indispensability, elimination of competition in a downstream market…) which makes it difficult to think of instances where it could be satisfied;
(v) I also observed that the main issue where competition law and data protection policies may converge relates to data portability. In cases where it is shown that scale is of the essence, then practices that could deny rivals a minimum viable scale could fall within the scope of the competition rule (in fact, Google’s proposed commitments -see here and here- already incorporate a section on the portability of data for AdWords campaigns). On the regulatory front, the proposed new EU regulation on data protection (currently stuck at the Council) also incorporates a right to data portability. Btw, some of the major companies cited in these discussions already have tools to facilitate portability (see here or here)
(vi) My last comment on this point was that privacy policies can also be a parameter of competition (even if admittedly many users currently appear to confer more importance to other parameters).
Apologies for making it so schematic, but having quite some work to do I’ve chosen to basically to a transcript of my notes, plus this is already lengthy enough for a post.
On the next post I’ll state my views on whether non-economic privacy considerations should be included as part of the consumer welfare standard.