Gatekeeper platforms in the EU Digital Markets Act (Part 1/2)
In June, the EU Commission launched two parallel public consultations for legislative initiatives to complement the existing competition law framework: one for a new market inquiry tool and another one for a regulation aimed at “large online platforms with significant network effects acting as gate-keepers in the European Union’s internal market.” Recently, Vice-President Margarethe Vestager announced that the two initiatives have been merged into one, which will be called the Digital Markets Act.
Booking.com has been supportive of the Commission’s competition policy agenda from the very start. In our response to the two public consultations we articulated our strong support for a broad market-structure based New Competition Tool and a targeted gatekeeper regulation addressing recurring anti-competitive behavior.
As we are getting on to the final stretch before the Commission’s proposal will be unveiled, the question of how to properly define gatekeepers takes center stage. In fact, it is crucial in achieving the Commission’s stated policy objectives of ensuring fair competition, and fostering innovation and competitiveness of the European online environment. A definition that captures the wrong set of companies risks leaving competition problems unaddressed and poses a severe threat to wrongly targeted companies’ ability to compete and innovate — in particular, if competitors are not facing similar restrictions.
This blog post offers our thinking on this matter.
Businesses must be dependent on gatekeeper platforms
While there are potentially many criteria for defining gatekeeper platforms, at its core a gatekeeper is characterised by its enduring ability to control access to a critical number of consumers. Businesses must be dependent on such a platform in the sense that otherwise they would be unable to reach a significant number of consumers to have a viable business. For such a dependency to arise, there must not be suitable alternatives to the gatekeeper platform at present and competitive alternatives must be unlikely to arise in the short or medium run because of considerable barriers to entry.
Such an approach is strongly supported by economic theory. For example, in his seminal paper on two-sided platforms, renowned Oxford University economist Mark Armstrong defines a “competitive bottleneck” as a platform where one side single-homes while the other one multi-homes. This gives the platform monopoly power over the multi-homing side and thus an ability to employ anti-competitive strategies. It is this significant market power that gives rise to concerns.
The logic is intuitively clear. If (business) customers have a choice, unfair behavior is unsustainable. If a platform mistreats them and there are alternatives, customers will just take their business elsewhere. If there is choice, platforms have to keep their customers happy. So the question with regard to defining gatekeepers is whether at least one side of the platform is locked-in (for whatever reasons) and can be reached only via the platform.
Lack of multi-homing as the defining characteristic of gatekeeper platforms
Putting customer lock-in at the heart of the gatekeeper definition is strongly supported by the EU National Competition Authorities (NCA). In their feedback to the consultation for a New Competition Tool (NCT), the NCAs characterised gatekeepers as follows:
“Gatekeeper scenarios refer to situations where a company controls the access to a number of customers (and/or to a given input/service such as data) who — at least in the medium term — cannot be reached otherwise. Typically, customers of gatekeepers cannot switch easily (‘single-homing’). […] [G]atekeeper scenarios are characterised by a lack of multi-homing, which can lead to significant market power, even in the absence of dominance.”
The lack of multi-homing was also the most important defining feature of gatekeepers named by stakeholders during the public consultation for the NCT.
Operationalising a definition according to these considerations thus requires at least three elements:
i) a platform is necessary conduit to a critical number of users,
ii) a platform needs to show a low degree of multi-homing on at least one side of the platform,
iii) there must be formidable barriers to entry. If all three criteria are fulfilled, a platform could rightfully be designated a gatekeeper.
To assess i) one might define some quantitative criteria and thresholds, e.g. turnover, traffic or consumer attention share. ii) and iii) would require a more qualitative evaluation with possibly some quantitative guidance on what would be considered a low degree of multi-homing. There might also be other considerations, for example a threshold to exclude niche platforms. However, the defining core needs to have the three elements above to target those platforms that can rightly be considered to raise concerns.
Such a three-pronged definition is also fully aligned with recent research of the EU Observatory on the Online Platform Economy (OOPE), carried out on behalf of the EU Commission in preparation for its legislative initiatives. In its report on measurement and economic indicators, the OOPE makes the following observations on measuring business dependence on platforms:
“One approach to assessing how dependent businesses are on platforms would be to measure how large a share of the businesses’ revenues are coming through a platform or platforms, as opposed to the businesses’ own websites and brick-and-mortar sales channels. For instance, if businesses in a given vertical or sector (e.g. hotels) get 50% of their revenues through a platform, then this strongly suggests that the vertical may be considerably dependent on the platform. However, it is not a definite indicator, because in theory the businesses might be multi-homing to maximize sales or they might have cheap substitutes that they are simply choosing not to use for the time being.”
“A complementary approach would be to measure “multi-homing”, or how many other platforms/channels the businesses are simultaneously using, regardless of their current revenue shares. For instance, if businesses in a given market use on average three other channels besides the platform in question to sell their products, then this suggests that they may not be very dependent on the platform, as they could switch the weight of their sales efforts to the other channels if threatened by the platform.”
“In a situation where a platform appears to be in a dominant or gatekeeper position, its economic power could in theory still be circumscribed by the competitive pressure coming from the existing market operators or threat of entry by new entrants.”
While the OOPE report provides valuable quantitative guidance on assessing dependency, it is also clear that designating a platform as a gatekeeper presupposes a careful analysis of the available data and market environment.
A need for quantitative and qualitative criteria
While there are some advocates — for example, for reasons of legal certainty — for employing purely quantitative criteria, the nature of gatekeeper platforms and in particular the need to assess the degree of multi-homing for a given business make such an approach inadequate. This does not mean it has to become arbitrary. But it does require more than a passing glance at some numbers and pre-defined thresholds. Such a more case-by-case approach could be complemented by clearly defined safe harbors to provide businesses with some certainty. For example, it has been argued that a platform that can show multi-homing on both sides of its platform should be exempted from a gatekeeper regulation. This is fully in line with our reasoning put forward here.
Getting the gatekeeper definition right matters
We have argued in this blog post that the defining characteristics of a gatekeeper platform is its grip on a critical number of consumers on at least one side of the platform. This creates a dependency of the other platform side on the platform for reaching these consumers. We have further argued, in line with economic theory and the observations from the National Competition Authorities, that this lock-in can best be assessed by analysing the degree of multi-homing present on both sides of the platform. The Observatory on the Online Platform Economy has provided valuable guidance on possible data points and thresholds for such an analysis and the EU Commission can and should build upon these.
We want to close this post by stressing that getting the definition for gatekeeper platforms right matters. Targeting the wrong platform and restricting its ability to compete on the merits, in particular while direct competitors would not face similar restrictions, carries a serious risk creating irreparable harm to competition, and ultimately consumers. We recognise that it is important to have a definition that does not create any loopholes for gatekeeper platforms to escape scrutiny and regulation. However, we wholeheartedly agree with the recent joint French-Dutch non-paper on gatekeeper platforms that “designing a regulation that does not exceed what is necessary to address the aforementioned concerns — i.e. which does not impose unnecessary obligations to the platforms without any gatekeeper position — is also important.” Getting the definition right matters, and is possible.
