First speaker of the second part was Miroslav Jakub, a Ph.D. student from the Charles University in Prague and the Parliamentary Institute, who presented a contribution dedicated to the search of right degree of trust in digital markets. In his contribution he explored reasons for digital regulation, compared Garman solution with Digital Markets Act (DMA) of the European Commission and provided conclusions. Among reasons for digital regulation Jakub mentioned complications related to economies of scale or privileged access to the platforms with data from different resources (and related economies of scope). Reasons were well demonstrated on the case Google Shopping (AT. 39740) in which the service Google Shopping was guilty for placing competing services on worse positions in the search mechanism. The case led to an extensive study of algorithm and Commission explained why some tests are not applicable (Bronner decision). Despite the fact, it is not easy to find a test which will be applicable to future cases. Recovery measures were very general (Google can not continue in this unfair practice). To sum up, the regulation shall be based on fast investigating mechanism and reaction, which is complicated as there is no proper test. This causes doubts when to intervene. Moreover, there are no tools how to approach full recovery. As pointed by Jakub, there are two ways how to look at it: 1) to give some arms to the competition law or 2) to create “right” ex ante regulation.
Miroslav Jakub presented German solution in the form of Gesetz gegen Wettbewrbsbeschränkungen – GWB) which is in force since January 2021 which is particularly interesting as it is trying to address online platforms. GWB is transposition of the directive 2019/I and gives special powers to Bundeskartellamt towards companies with “significant degree of market permeation”. In other words, the law is creating special type of the addressee based on criteria slightly different then DMA but aimed in the same direction. Moreover, there are some duties according to § 19a GWB – for example a ban to prefer own offers in relation to providing access to the market with products. However, efficiency defence is still applicable and there are many more interesting aspects of the law, including preventive measures to avoid “tipping” or introduction of “judiciary fast-track” for complaints. Miroslav Jakub excellently compared GWB with DMA proposal and concluded, that GWB is much closer to “classical competition law” while DMA is closer to independent and parallel ex ante regulation. Jakub concludes, that in favour of DMA is legal certainty, attempts to create equal conditions of the guardians and foremost the speed. However, German regulation is seemingly allowing targeted interventions against well established actors, offers greater potential for innovation and in comparison, to DMA might have be grater incentive for judiciary control.
Second speaker was Simona Rudohradská, from Pavol Jozef Šafárik University in Košice (Slovakia), who in her contribution looked at the issue of Gatekeepers according DMA in a broader perspective. After the broader introduction to the documents of the European Commission related to digital single market, she focused DMA and provided many interesting details of this new regulation. For example, in the proposed regulation there were three ways how to define big on-line actors: 1) to crate a list of “Gatekeepers” with set duties; 2) to create a partially flexible framework for identification with set duties for gatekeepers matching criteria (which was finally selected); or 3) to create flexible framework based on qualitative values of scope. As a result, there is closed list of platform services providers (article 2 DMA), there is also a combination of quantitative and qualitative criteria (Article 2), directly applicable duties (Article 5 and 6) and options for the Commission to update tools regarding the duties. Among criteria is important the influence on the internal market, character of the activities and permanent position of the activities. The criteria are further specified based on total revenues etc. The duties of gatekeepers are starting with notification of the Commission and providing required data. Even when failing this duty Commission has right to indicate status of gatekeeper. Among other duties, gatekeepers shall prevent connection of personal data from basic services with other personal data from different platforms or services, they shall allow commercial users to offer same products or services to end users via online services of third parties for prices or conditions that are different from prices or conditions offered via on-line services provided by the gatekeeper. Also, commercial users shall not be prevented from reporting authorities any issue related to gatekeepers or any restrictions made by gatekeepers.
From the theoretical perspective, Simona Rudohradská pointed out that current unfair practices of gatekeepers are punished according to the existing regulation. DMA is merely complementary and shall not touch existing tools at the national level. The DMA proposal is positive step and merely a “superstructure”. Creation of national norms might cause problems as gatekeepers operate internationally. That is why selected way of gatekeeper identification and mechanism for setting the duties is flexible and appropriate to the aims of regulation. Moreover, identification based only on quantitative criteria can lead to omission of relevant actors with specific influence or position.
In the following contribution Martin Abraham, an advocate with specialization on competition law, asked a simple question: are tools of competition law in the digital age sufficient? In his point of view Europe is far behind in the regulation as US based companies such as Apple, Alphabet, Amazon or Facebook reached dominance in their areas. Next to the competition issues there is also a geopolitical dimension. There are some tools, which are available, but solving situation ex post – addressing issues such as dominant position, cartels, or merges. However, in the case of cartels the tools are limited and regarding merges the case of Facebook and WhatsApp shows, that without notice of authorities there was a merge strongly empowering Facebook. That is why Martin Abraham is for EU regulation and thinks that Germany shall use its political power in Brussels to push for a regulation which will be not duplicated at the national levels. Diverging regulation for small countries is politically not sustainable as shows the clash between Google and Australia after the introduction of a special fee for using media. The solution shall be the DMA addressing the issue of gatekeepers who are having strong position, are operating in several countries and are having real impact on the market. As for DMA key aspect is interoperability allowing access for small companies (underdeveloped actors from Europe) and access to the data for business. According to Abraham, this form of regulation is required and required are also restrictions according to the DMA. Regarding the tools, there is a market investigation and sanctions, allowing to fine gatekeepers up to 10 % of total turnover or repeated 5% of daily turnover. However, Dr. Abraham was surprised, because the tools are not so powerful as in the UK regulation, enabling Commission to require structural changes. To sum up, he is not a friend of robust regulations, but this one is necessary and he warned from adopting divergent regulations at the national level. As highlighted by the moderator, very interesting question regarding sanction is the principle ne bis in idem and the applicability of punishments at the national and EU level, which is according to Dr. Abraham not mutually exclusive.
Another speaker was Mária T. Patakyová from the Komenský University in Bratislava, who delivered contribution about price algorithms in the context of EU competition law. After introduction of the article 101 TFEU and main attributes of unfair practices she moved to algorithms. Regarding the agreements thee is the necessity of being 1) at least bilateral and 2) informal but with elements of acting not independently. It means that acting is not natural and may be obvious from the acting on the market – for example exchanging sensitive information, setting prices at higher level etc. As pointed out in case C-1999/92 P Hüls, there must be causal nexus in such activities. Agreements can change in time into concerted practice and is not mutually exclusive. The core aspect is the aim or relevant manifestation as exclusivity, restricting competition or its distortion, usually leading to lower level of economic activity, negative impact on the market or different impacts. Mária Patakyová opened very interesting question whether a silent collusion is escalated by algorithms and pointed out that there are two opinion steams: that situation is relatively stable or getting worse. She pointed out that algorithms are for several years debated in connection to Messenger, Hub &. Spoke, Predictable Agent and Digital Eye which is using autonomous algorithms.
After detailed introduction she has selected price algorithm which was developed by a third party (software company), which is bought by fictional companies on a market, where are just several actors and high barrier for entering the market. At the same time there is no written agreement among companies that it is necessary to buy this algorithm. Benefits are related to sensitive information exchange, higher price on the market. The practice was demonstrated on several variations. Variation 1 is horizontal agreement about purchase of the algorithm and price-fixing – which is not a problem. In the variation 2 there is no agreement, we can talk about concerted practice in case of same behaviour on the market. However, to prove is complicated and may be proven by contacts in between companies and avoiding another cause. However, in the variation 3, without previous contacts of the companies the same behaviour on the market will be not punishable as a concerted practice (which is confirmed by OECD in 2017). There is also Variation 4 involving vertical agreement based on block exception and specific situation on the market. As a result, application of the article 101(1) TFEU has its limits to cover all situation where silent collusion is present. When companies will justify the purchase of the algorithm, they can avoid application of concerted practice and punishment. As noted by the moderator, great challenge are self-learning algorithms which might be bought without knowledge of the purchaser. As pointed by Patakyová, it is a question of time when such algorithms will enter small markets and cause the increase in prices. Michal Petr highlighted also technical boundaries of algorithms being on the edge of understanding for lawyers. (End of part II.)
The event was organised within the implementation of the Jean Monnet Network „European Union and the Challenges of Modern Society (Legal Issues of Digitalization, Robotization, Cyber Security and Prevention of Hybrid Threats) Project id: 611293-EPP-1-2019-1-CZ-EPPJMO-NETWORK.