On the Amazon probe: neutrality everywhere (or the rise of common carrier antitrust)
Commissioner Vestager announced, last week, the launch of a preliminary probe into some of Amazon’s practices. According to the information that has been made available, the probe has been triggered by Amazon’s use of data coming from merchants using its marketplace.
Because the users of the marketplace are also Amazon’s competitors, the practice would create an uneven playing field that would be tilted to Amazon’s advantage (Amazon would benefit from all the insights coming from the data). As far as I can gather, this is what the case is about.
As soon as the probe was announced, I received a call from Bloomberg’s Aoife White (see here). I told Aoife that this trend seems to mark the rise of what I called ‘common carrier antitrust’ (the concept did not make it, alas, into the piece).
What do I mean by that? I mean that vertically-integrated online platforms are expected not to discriminate against rivals that are also their customers.
The wind seems to be blowing in a clear direction: we may be reaching the point where it is problematic, in and of itself, that an online platform’s affiliate is treated more favourably than third parties. This trend has been going on for a while: it is not by chance that Neutrality Everywhere? was the topic of our 2nd Chillin’ Conference.
If discrimination indeed becomes an antitrust problem in and of itself, online platforms may soon be subject, across the board, to the same sort of access, non-discrimination and accounting separation obligations to which incumbent telecoms operators are already subject (i.e. common carrier).
Is common carrier antitrust new? Is it business as usual?
This trend, if it wins the day, will change the shape and scope of EU competition law. So in short: no, I do not think it is exactly business as usual.
- In essence, common carrier antitrust involves the extension to all dominant companies of some principles that originally applied in a relatively narrower context.
- Strict common carrier obligations make sense in the context of Article 106 TFEU (companies that benefit from exclusive or special rights). GT-Link is a wonderful example in which strict non-discrimination obligations are uncontroversial.
- It is not unreasonable to extend common carrier principles to dominant companies that control an indispensable input (as in Commercial Solvents and Telemarketing). This is a point that was also made by the Court of Justice in Deutsche Telekom (see paras 230-233).
- The Amazon probe is one of many recent examples that signal the likely expansion of common carrier obligations in relation to non-indispensable inputs – Google Shopping is probably in the mind of everyone reading this; but Android, at its heart, is also about a vertically-integrated firm favouring its affiliates through various contractual devices. The Internet Explorer case was closed in a very ‘common carrier’ way: a must-carry obligation placing Microsoft’s and competing web browsers on an equal footing vis-à-vis the platform (operating system).
- Amazon might (only might, we do not even have a formal investigation) signal a step forward in the expansion of common carrier antitrust. If discrimination is deemed problematic in and of itself – that is, absent actual or potential foreclosure, then the practice would have become prohibited by its very nature (that is, by object). The same would happen, for all practical purposes, if the threshold of effects were set at a low level.
- If so, the case would mark the expansion of common carrier antitrust all the way through the legality spectrum: the scope of common carrier antitrust would have extended from Article 106 TFEU firms –> to indispensable inputs –> to dominant firms (subject to an effects analysis) –> to dominant firms (by object).
If we accept that there is, in some respects, a friction between common carrier antitrust and the case law, the next question is whether this trend, under which discrimination may end up being perceived to be a problem in and of itself, is desirable.
My view is that we should have an open and explicit debate about the merits and consequences of common carrier antitrust (for competition law and for markets), and about the range of legal tests that can be used to address discriminatory conduct. It would be a bad idea to just pretend that it is business as usual.
In this regard, I guess my point of view is well known:
- EU competition law never saw vertical integration, or favouring an affiliate in a vertical integration setting, as a problem in itself. On the contrary. It has long been accepted as undisputed that discrimination by vertically-integrated firms is not only pervasive but is often actually pro-competitive.
- In many instances, discrimination (which manifests itself in many ways) paves the way for competition; sometimes, it is what makes competition possible in the first place. Some products and some innovations exist only as a result of a vertically-integrated firm favouring an affiliate.
- I can think of many examples, some of which I discussed with Aoife:
- HBO, which has transformed TV for the (much) better, is a child of discrimination. Cable companies expected to attract subscribers to a nascent technology by keeping attractive programming to themselves, and by giving it more favourable treatment within the platform.
- Supermarkets favouring their private labels has not only done wonders for the competitive process, but has also brought prices down and improved the shopping experience of ordinary people.
- I have discussed the example of the Tour de France on the blog. The Tour de France was created by a newspaper to attract new readership. The point of creating the event was to favour its affiliate so it could give better and more exhaustive coverage.
Some people may argue that the Amazon probe is addressing Amazon’s incentive to favour its own service. Commentators have spoken, in this regard, of a ‘conflict of interest’ (that is, an incentive to foreclose), which would be the real problem (see here, for instance).
Again, the said conflict of interest cannot be – at least, it has never been – a problem in and of itself. And this in spite of the fact that the incentive to foreclose will in many cases be real and difficult to dispute.
Why do firms refuse to license their intellectual property rights? Because they have an incentive to keep the downstream activity for themselves (i.e. they have an incentive to favour their own downstream services over those of rivals). There is probably not a much better example of a conflict of interest.
Taken to its logical consequences, the claim that the conflict of interest is in itself a problem, would mean that dominant firms should not be entitled to refuse to give access to their own tangible or intangible property.
And, of course, that would be a major departure from competition law as we know it – if there is something that we know for sure is that, no matter how undisputed the conflict of interest, only in exceptional circumstances are firms required to license their intellectual property under Article 102 TFEU.
***
I guess the many words I have written above can be summarised fairly easily: common carrier antitrust would entail, in some respects, a departure from EU competition law as we know it.
If this is the direction of travel, let us have a long deep thought about the major consequences that come with this new way of conceiving competition law. In this sense, the consultation launched by the European Commission on this question comes across as the right thing to do (and note that you still have time to submit your views).
Chillin'Competition 
This assumes that Amazon is not indispensable while, for example, Royal Mail’s bulk mail business is: https://www.ofcom.org.uk/about-ofcom/latest/features-and-news/royal-mail-whistl-competition-law
That seems, to put it mildly, a claim that could do with further evidence.
Martinned
25 September 2018 at 11:12 am
As ever, Martin, thanks for being a faithful follower of the blog
There is a chance that you missed the word ‘likely’ in my analysis. Why likely? Because the Commission may well claim that the input is indeed indispensable, and because showing indispensability is very hard (and we have plenty of experience to show for it).
More generally, your suggestion that indispensability has to be disproved, as opposed to proved is fascinating (as well as a delightful attempt to shift the burden of proof).
We know indispensability is a truly exceptional occurrence. Google is an immensely powerful company and one of the major gateways to the Internet. Even in that case, the Commission fell short of claiming indispensability.
Arguing that Amazon is indispensable for merchants comes across as an extraordinary claim that would need to be proved, to the requisite legal standard, by an authority. Plausible claim? Perhaps. But there is no reason to presume indispensability in relation to Amazon. Suggesting that Amazon’s indispensability can be presumed and thus has to be rebutted by the company is even more extraordinary.
Pablo Ibanez Colomo
25 September 2018 at 11:30 am
It might be even harder to prove if, according to the Commission’s own findings in the sector inquiry, a mere 4% of retailers rely only on marketplaces for online sales. Having re-read section 4.4 of the SWD, I very much doubt the Commission would even try to show indispensability.
That said, there are judgments which the Commission could rely on to avoid the indispensability test. An obvious candidate would be the 1997 Deutsche Bahn / Transfracht judgment, which essentially lays down a common carrier standard. A co-founder of this blog once relied on that judgment in a dispute with a former President of the GC 🙂 As he pointed out there, that judgment may be unprincipled, but it still part of the body of law.
As to whether common carriage for all dominant companies would be desirable… there’s a book by Jeff Malcolm of NERA in which he illustrates that it doesn’t even work for gas pipelines. Difficult to imagine that it could be an efficient form of dealing with replicable assets which just happen to be controlled by a company with market power.
HK47
25 September 2018 at 6:49 pm
I’m sorry Pablo, I didn’t mean to shift the burden of proof. Obviously indispensability is for the competition authority to prove. I just meant to put the line for indispensability somewhere other than “even Google isn’t indispensable”, as demonstrated by the example of Royal Mail. It’s surely easier to set up another bulk mail delivery system than to set up another Google or Amazon. (As, respectively, Microsoft and Facebook are discovering as we speak.)
The fact that the Commission hasn’t (yet) claimed that Google is indispensable seems irrelevant. They approach their cases in whatever way makes it easiest to win in court.
Incidentally, the number or share of retailers who rely on Amazon is surely irrelevant, unless you mean to define “indispensable” as “indispensable for everyone”.
martinned
25 September 2018 at 10:56 pm
Thanks both for your comments!
Martin: it is not immediately obvious to grasp what you mean by indispensability. What seems clear, in any event, is that your interpretation departs from the definition given by the Court in IMS Health. The evidence at which HK47 hints would be conclusive under that definition.
HK47: Deutsche Bahn is a case that falls within the GT-Link/Deutsche Telekom continuum. As such, it only confirms my point. As you can understand, I cannot be held responsible for what other people may have tried to make the judgment say!
What is interesting is not so much to play around judgments and stretch their meaning, but the issues you raise in the third paragraph. If confronted directly with the question, the Court will have to decide whether common carrier obligations apply, as a matter of principle, to all dominant firms that are vertically-integrated. No statement to this effect has ever been made. The implications are substantial and would change EU competition law, and this, for the very reasons you identify.
Pablo Ibanez Colomo
26 September 2018 at 7:28 am
Pablo, I understand your comments revolve around competition laws but maybe thinking outside the box would enrich the debate. Competition rules are worth for what they are worth (in my opinion, basically cartels and making sellers’ life difficult with biased vertical restrictions).I would love to see your take of this and other cases from the angle of unfair competition rules. For example, you/consumer may be very happy with the supermarket brands that, at first sight, bring more variety and lower prices. But you may be missing the backstage abuse: a supplier providing advance info of innovations and the retailers using it in favour of their own brands, no surprise DG COMP’s 2014 retailer concentration study found a decline in innovation as of 2006) copycat packaging, switch marketing in the store, shifting retailer risks onto suppliers (see the Edeka case of the BKT in Germany) and so on so forth. By the way, I am more confident on our ability to identify an unfair practice than an anticompetitive practice (leaving aside cartels), despite all your efforts to bring some clarity into it. A law that is predicated on a parameter – consumer welfare- that exposes inherently contradictory aims – innovation/quality and prices) will always be uncertain, despite the authorities’ flawed focus on low-prices.
Thanks for all your hard thought-provoking work
J&B
26 September 2018 at 9:28 am
Thanks, Justerini!
It is great to get an alternative perspective. And it would be wonderful to chat more extensively about these issues, which I find fascinating.
I am not sure that unfair competition law makes things more straightforward.
I, for one, fail to see why the practices you describe are unfair. I would call them ‘competition as usual’. If supermarkets do not infringe the suppliers’ intellectual property, what is unfair about copying the suppliers’ innovations? Would you say it was unfair for HTC, Samsung and others to manufacture a ‘copycat’ of the iPhone pretty much as soon as it was released?
I fully understand the anxiety of some suppliers, but their anxiety should not necessarily be a reliable indicator of lack of fairness, or of harm to the competitive process (and note that I distinguished between harm to the process and harm to consumers, as there is no clear statement in EU competition law that the latter is the aim of the system in the EU).
Deep insights in any event. Thanks again!
Pablo Ibanez Colomo
26 September 2018 at 2:37 pm
Well, the distinct and unfair feature here is that suppliers must share with their buyers/competitors their trade secrets (innovation launches and all the supporting plans&surveys) months ahead and buyers/competitors can use these trade secrets from day one. A trade secret breach may be an issue here (misuse) but no one challenges a big buyer. To take your example: assume that Apple was “forced” to share its innovation plans months ahead with HTC, Samsung and others in order to be able to market its Iphones and Apple cannot sue them for fear of being cut off from consumers. The point here is that many unfair practices that would surely be considered abusive if committed by a dominant company are committed by the leading operators in the market. The market disturbance seems to me to be much more intense when, say, 4 non-dominant operators with a combined market share of 80% rely on these practices as compared when these practices are committed by an operator holding a 50% market share (assuming it is dominant).
Brooks 😉
J&B
27 September 2018 at 10:45 am
I’m not sure whether unfair competition provides a more coherent framework for assessing anything. In the German-speaking world, which has a fairly developed tradition in that field, you have precedent for any outcome you may desire, as well as for its exact opposite. Applying this in practice is an art, and I have great admiration for my colleagues’ skill in making their opponents look awful and their clients like saints. But there is a reason why it is applied by civil courts and cease-and-desist orders, rather than by deterrent fines. It may be a mess, but at least we minimize error costs…
HK47
26 September 2018 at 3:12 pm
[…] On the Amazon probe: neutrality everywhere (or the rise of common carrier antitrust) Commissioner Vestager announced, last week, the launch of a preliminary probe into some of Amazon’s practices, apparently triggered by Amazon’s use of data coming from merchants… Pablo Ibañez Colomo (Chilling Competition) […]
Blog O Blogs September | Competition Policy International
3 July 2019 at 3:27 pm
[…] online platforms has been an issue of modern competition law, inter alia, in the EU. According to Pablo Ibanez Colomo, a law professor at the London School of Economics, the Amazon case is just another example of the […]
European Commission opened a formal investigation into Amazon’s data practices | Sangyun's Blog
26 July 2019 at 10:12 pm