Responding to digitization in the German Competition Act
By Dr. Michael Holzhäuser and Nadja Waksman
Continued legal reforms are necessary to keep up with a rapidly developing digitized world that is intertwined with data protection and privacy. In investigating abuse of “dominance” (market power) cases, old parameters, such as market shares and barriers to entry, no longer provide adequate tools to truly capture a data-intensive company’s position in an inherently dynamic market. With data being the new raw material of the future, regulators must find a delicate balance between under- and overenforcement to protect innovation in markets where entry can easily be disruptive. Further, new tools are needed to deal with the economics of multisided platforms.
Germany has sped ahead with a further draft amendment (Novelle) to its Competition Act (GWB), having already adjusted its merger notification thresholds to capture the transaction value of deals. German competition legislation has also introduced the concept that a “market” need not be linked to monetary compensation. The Novelle focuses on the challenge of “access to data” as an indicator of market power, countervailing the tendencies of monopolization of platform markets and the position of intermediaries in multisided platforms. This legislative change is accompanied by a call by the German Competition Authority (Bundeskartellamt, BKArtA) for more competencies, particularly in the areas of consumer and data protection.
At the time of this writing, other jurisdictions (such as Australia, France and Singapore) have focused on carrying out market reviews and setting out recommendations, with some countries being satisfied that their current legislative tools are sufficient to deal with the challenges of a digital economy. The UK government has indicated that recommendations set out in the Furman Report will be included in a competition green paper later this year and the JFTC in Japan has indicated that “unfair means of personal data collection.” and “unfair utilization of personal data” could be new types of abuses. Responding to changes in a digitally transformed economy also calls for a reform on a European level, including an evolvement of Article 102 of the TFEU. The Commission’s response to the “Competition policy for the digital era” report and the G7 summit this summer remains eagerly awaited.
Background to the Novelle
The German Federal Ministry of Economics and Technology intended to address both German and European platforms with its latest draft amendment to the GWB (as of October 7). The Novelle will be presented to other ministries for their approval before being laid before the German Parliament.
The amendment runs in tandem with the recent industry report titled “Commission Competition Law 4.0” and is designed to implement the ECN+ Directive increasing the efficiency of competition law. Dealing with data in defining market power, broadening the scope of abuse of dominance theories and granting better consumer protection form the backbone of the Novelle. It also echoes a study by Haucap, Kerber and Schweitzer, which was commissioned in 2018.
The amendment should also be seen against the backdrop of recent EU Commission decisional practice that emphasizes the low “capability to restrict” standard in abuse of dominance cases and novel theories of harm. For example, the infringement decision in Google Shopping was based on a theory of “self-preferencing” that was deemed capable of excluding rival comparison shopping sites. In that sense, the EU Commission is sanctioning a “new type” of practice.
Data as indicator of market dominance
To demonstrate abusive behavior, it is first necessary to prove an undertaking was dominant (had market power). The Novelle contemplates introducing the concept of “access to data,” which is a new parameter in assessing dominance alongside established factors including market shares. The broadening of the concept of market power is significant, as it is likely to apply to a large group of data-driven businesses including social media platforms, fintech businesses and online booking sites.
Businesses will now need to have a good grasp of the volume and type of data they hold and to what extent their data set is unique and if, how, where and which other market operators would be able to access a data set from other sources. Not only is greater vigilance needed to avoid a potential abuse of such market power (such as unlawfully “restricting access to data”), but the wider definition of dominance will also influence how data-intensive mergers are assessed where incumbents may have a competitive advantage that is not reflected in their market shares.
The importance of an intermediary’s service as indicator of market dominance
Further, in the context of multisided markets, the importance of an intermediary’s services in accessing supply and sale markets has to be considered in determining a company’s market power. The increasing presence and growth of online multisided platforms (such as Uber, Airbnb and online travel agents) have long prompted the need for greater intervention.
The 2018 study by Haucap, Kerber and Schweitzer introduced the concept of “intermediary power” as a criterion of market dominance. The more an intermediary can bundle the demand for products, the more the suppliers of such products will be dependent on the platform for access to the customers. This dependency can exist even below the current thresholds for market dominance, as the BKArtA following the contemplated legislative wording will now have the power to conclude that an undertaking in a multisided market enjoys a position of “paramount importance.” This language mirrors the notion of digital platforms with “strategic market status” in the Furman Report. One of the factors the BKArtA will look at in determining this status is “access to data.”
Refusing supply of data: a new type of abuse
Applying traditional theories of harm to “digital problems” while preserving legal certainty is arguably the main challenge that legislators face. Although the list of potential abuses of dominance is not closed, there are clear difficulties in mandating ”access to data” within a predictable legal framework.
Today, an undertaking would have to offer access to data if it has a dominant position and if that data was deemed “essential” or “indispensable.” However, EU Competition Commissioner Margrethe Vestager stated earlier this year: “Access to data has to be redesigned so that newcomers can compete with big tech giants.” The Novelle has attempted to redesign access to data. Under the proposed amendment, a dominant undertaking may be liable if it refuses to grant access to data, data networks or other infrastructure (including its own) where such supply of data is “objectively necessary” for a market participant to operate in either an upstream or downstream market. This is subject to the caveat that such a refusal would harm effective competition and there is no objective justification for such a refusal.
If this concept is ultimately adopted in the GWB, it remains to be seen how “objectively necessary” will be interpreted by the courts in Germany and how much lower a threshold it will be compared with the current “essential facilities” standard. However it may be applied, this is a remarkable change, as it broadens the traditional “refusal to supply” theory of harm; it may require access to data without compensation (according to the commentary in the amendment); it removes the defense that supplying data would be impossible or unreasonable to expect; and the duty may be applied to businesses that lack market power in the traditional sense (given the broader data-based definition of dominance described above).
Applying the proposed amendment could mean that it would be unlawful for a bank to refuse fintech companies access to customers’ banking data if the bank is the sole owner of that data and provided that customer consent has been granted. Banks arguably have an incentive to foreclose fintechs when they offer competing products based on bank-account data — an area to be watched closely, particularly if companies like Facebook and Google enter the financial services space.
The requirement to obtain consent also highlights the jurisdictional conundrum of which enforcement body should be responsible for determining that access to data should be granted — in other words, is it more a competition or a data-protection issue? Although the recent decision by the OLG court in Düsseldorf could be heralded as a victory for Facebook, it does show that it could be appropriate for the BKArtA as a competition authority to investigate a privacy matter given that an exploitative theory can accommodate privacy-related issues. In that case, the BKArtA had, however, failed to show what the competitive level of privacy would have been, which shows one of the many difficulties of competition authorities assessing privacy matters.
Access to data has moved to the forefront of the debate surrounding the digital economy. EU officials have recently suggested that providing access to data could also be a form of remedy to restore competition, even after a business has been found guilty of abuse.
New types of abuse in multisided markets
As explained above, the BKArtA has the power to determine that an undertaking enjoys a position of “paramount significance for competition across markets,” that is, an intermediary that has a special strategic position — this a separate question as to whether an undertaking is dominant. In fact, rule 2 below shows that these questions are treated separately. If so, some types of behavior may now infringe competition law, such as:
1. Treating offers from competitors differently from its own offers (of services or products) when providing access to supply and sales markets
2. Shutting out competitors to the detriment of competition where the undertaking enjoys a competitive advantage and can expand rapidly (even in the absence of any market power)
3. Creating barriers to entry in a dominant market by “free riding” on data that has been collated by another undertaking or demanding terms and conditions permitting such use of data
4. Making the portability of data more difficult and thereby preventing competition
These types of conduct are subject to the traditional (and notoriously difficult to prove) exemption that a restriction was objectively justified, which the company accused of any wrongdoing must demonstrate.
The rule against self-preferencing or discriminating against competitors’ offers (rule 1) may be particular stringent, as it is designed to capture undertakings that offer goods or commercial services on a platform on which other undertakings strongly depend and where consequently a strong asymmetry exists. Businesses cannot therefore prefer their services to those of competitors (Selbsbevorzungsverbot).
Applying this rule means that platforms such as Amazon, for example, would have to ensure that they do not discriminate against offers from other merchants where a platform has a dual role of providing its own products as a merchant and advertising those of third-party merchants.
In response to the BKArtA’s investigation into Amazon, Amazon changed its terms of business for sellers on its online marketplace. Despite this settlement, Amazon remains under scrutiny, particularly as small and medium-sized businesses in Germany argue that they are being squeezed out of the market — a sentiment echoed by some political parties and trade associations in Germany. Their main concern is that small and medium-sized businesses increasingly depend on dominant digital players.
Linked to this is rule 3, which is designed to prevent dominant undertakings from using data collated by third parties to their own advantage. One way of preventing this would be to create Chinese walls, although it begs the question how this could be practically achieved.
Rule 4 enhances consumer protection by increasing data portability — a key concern voiced in the industry report. In practice, users’ search history could be transferred from one search engine to another, for example. Again, this raises considerable data protection issues.
Greater data portability also prevents the risk of “tipping” in highly concentrated markets as explained below. The industry report had highlighted the concern that platforms with pronounced positive network effects can lead to highly concentrated markets in which an undertaking can become a “gatekeeper.” Such effects can create high barriers to entry precluding the entrance of new market players (referred to as the “tipping effect” once entry into a market is no longer possible).
According to the industry report, this phenomenon is increased when a company benefits from customer experience on one side of a platform to grow on another side of the market. Facebook, to illustrate this with an example, operates a social network platform and online advertisements provided by third parties. The more consumers use Facebook’s network and hit certain “likes,” the more advertisements will be targeted at them.
Another way of reducing this risk is to encourage consumers to “multihome” (that is, to list services on several sites), which in turn requires greater data portability as the industry report suggests.
Mergers: some relaxation of rules, but no killer solution
One of the notification thresholds has been increased, which has arguably excluded some transactions that do not have a significant impact on the market. This is reinforced by the introduction of a de minimis clause that exempts transactions below a certain threshold.
Given the recent debate surrounding them, it is surprising that the Novelle is silent on “killer acquisitions” — that is, acquisitions of an innovative target company with the aim of ”killing” either the target product (as in cases in the pharmaceutical sector where an incumbent buys out a pipeline product belonging to the target) or eliminating future independent competition in that market (as is arguably more the case in the digital sector).
Identifying killer acquisitions in the digital sector is particularly problematic, as it not only requires an assessment of immediate harm but also an evaluation of how an embryonic competitor may have developed absent the merger. Given that it is hard to predict what the future effects on an existing or potential competitor are on a forward-looking basis, calls for an ex post review of mergers (that is, once the merger has completed) have become louder. However, the Novelle makes it clear that the current ex ante approach could remain in place, despite the BKArtA’s call for an ex post review.
One way of dealing with such acquisitions would be to capture the deal value of transactions. The EU Commission seems to have adopted a wait-and-see approach depending on how the BKArtA and the Austrian competition authority enforces their newly amended notification thresholds. A further difficulty raised by such acquisitions is who should bear the burden of proof: The EU Commission’s policy report referred to in the introduction of this article recommends shifting the burden to the incumbent to prove the pro-competitiveness of its conduct.
Burning questions remain unanswered. One way of dealing with killer acquisitions in the digital sector may be to focus on an innovation-based competition assessment as developed in the GE/Siemens Alstom, Dow/Dupont or Bayer/Monsanto strand of cases and the protection of “innovation spaces” as the CMA has suggested by the CMA in its consultation on digital mergers.
Guidance on restrictive agreements
A change to be welcomed is the introduction of a voluntary notification request for collaboration agreements that is reminiscent of the premodernization EU approach under which parties would have to first notify a competition authority about an agreement. This move will inject more legal certainty into the current approach in the spirit of collaboration between market players and the BKArtA, especially where agreements are complex. Parties will receive a decision within six months of submitting such a request.
Conclusion: some answers, many further questions
Breaking up tech giants completely (as increasingly suggested by Elizabeth Warren, the current Democratic front-runner in the US) was deemed to be a method of last resort by Vestager when she spoke at a tech conference earlier this year. However, more empirical evidence is needed to understand how easy market entry is where businesses face large incumbent digital players.
While the suggestions in the recent Novelle may be applauded as an attempt to deal with new challenges in a digital era, competition regulators and legislative bodies around the globe still face a number of to-dos on their list. Some of these are set out in the questions below. Not all of the industry report’s recommendations have been adopted in this Novelle — this includes the suggestion that platforms with a certain minimum level of users or sales implement an alternative dispute-resolution procedure using independent third parties to address infringements on platforms.
But perhaps we need to wait for the next wave of another Novelle …