First practical experiences with the private enforcement of the German ‘tipping paragraph’ (Sec. 20(3a) Competition Act)
In early 2021, the German legislator amended the country’s antitrust law and introduced, inter alia, Sec. 20(3a) to the German Competition Act. The provision is dubbed the ‘tipping paragraph,’ as its goal is to specifically reduce the risk of tipping in multi-sided platform markets. Within the first year of adoption, there have already been four noteworthy decisions applying the new provision in practice and guiding its future interpretation by the courts. This article provides an overview of the first practical experiences with Sec. 20(3a), with a particular focus on its application in private enforcement actions.
There is consensus in the economic literature that in multi-sided digital markets, platforms are more prone to reach a so-called ‘tipping point’ than in traditional markets of the ‘old economy’. This is notably due to the presence of positive network effects which often create a natural tendency for market concentration. If network effects are sufficiently strong and not effectively counterbalanced by other factors (such as multi-homing), users are naturally drawn toward the platform or network with the highest number of users. As a consequence, the market leader becomes more and more attractive – until the market eventually ‘tips’ into its favor and becomes a market with only a few operators or even a monopoly. Since this process is often irreversible, it presents a dilemma for policy makers: Should ‘tipping’ be prevented, and if so, under what conditions?
The German legislator opted for intervention. The lawmaker decided to extend the toolbox to prevent market tipping – or even the risk thereof – at an early stage. However, intervention is only possible if the market leader increases the pre-existing tipping risk by employing obstructive practices which depart from competition on the merits.
To this aim, the German Competition Act was amended in early 2021. Since then, a new Sec. 20(3a) prohibits unfair unilateral practices where (i) an undertaking with superior market power (ii) in a multi-sided market (iii) impedes the independent attainment of network effects by competitors and, in this way, (iv) creates a serious risk of significantly restricting competition on the merits. When such prerequisites are fulfilled, conduct is deemed to be an ‘unfair impediment of competitors’ and can be prohibited by (i) the Federal Cartel Office in the course of public enforcement or (ii) by courts in private enforcement actions of competitors.
‘Market tipping’, ‘network effects’ and ‘multi-homing’ as a starting point
As the German legislator already recognized in a previous amendment to the German Competition Act in 2017, multi-sided markets and networks are characterized by the existence of network effects which naturally increase concentration tendencies in the market. Notably, the presence of strong network effects can lead to what is commonly referred to as ‘winner-takes-all’ markets.
There are different types of network effects. Positive network effects arise between different users or user groups when the value of a product or service to a user increases with the number of other users. Conversely, they are negative if such value decreases as the number of users grow. Network effects are direct if they arise within the same user group (i.e., within a network) and indirect if different user groups benefit (as in the case of two- or multi-sided markets).
An assessment of market tipping naturally involves elements of uncertainty and must be based on probabilities, potential risks, and incentives. First, the process of market tipping is not equally likely in all multi-sided or network markets because the likelihood of its occurrence depends on a number of other, interdependent factors. According to the German legislator, “predicting when and for what reasons the tipping of a platform will occur is difficult”. Second, tipping is not only predetermined by the structural properties of a market, but can also be favored by certain practices of individual market players. Third, platforms have a considerable incentive to seek to reach the market’s tipping point due to the prospect of monopoly profits. Fourth, once a market has tipped into the lap of the market leader, it is practically impossible to reverse this process.
However, there is also a consensus in the economic literature that unhindered, widespread multi-homing can effectively counter the pre-existing, natural tipping tendencies in a market. Multi-homing, i.e., the parallel use of several platforms or networks, enables platforms that are competing with the market leader to generate their own network effects and remain or become competitive. Consequentially, tipping risk and multi-homing are inextricably linked.
Against this backdrop – and notably in view of the elements of uncertainty –, the German legislator amended the countries’ antitrust code and opted for an early intervention point. Due to the inability to determine the optimal point of intervention in each individual case, the lawmaker decided in favor of the application of a precautionary principle.
To overcome the gap in the protection of competition in the relevant markets that the traditional abuse of dominance provisions left, the new Sec. 20(3a) was adopted within the framework of the 10th amendment to the German Competition Act. The 10th amendment entered into force on January 19, 2021, and was tellingly entitled “Competition Law Digitalization Act,” as it contains numerous provisions aimed at safeguarding fair competition in digital markets. As such, the provision is in line with the general objective of the amendment, as to which the legislator was pursuing an overall acceleration of the competition law enforcement and an increase in its effectiveness given the dynamics in the digital economy.
While recognizing that the process of market tipping is not per se anti-competitive, the legislator made clear that any conduct which departs from competition on the merits, and which increases the pre-existing (i.e., natural, or abstract) risk of market tipping, is illegal if it creates a serious risk of a significant restriction of competition. However, while the German Government’s advisory committee recommended introducing such an additional prohibition only for markets which are at a concrete tipping risk, the legislator did not follow their recommendation as regards this higher threshold for intervention. In fact, the actual ‘tipping process’ can nowhere be found in the wording of the provision, and is, as such, not a prerequisite to finding an ‘unfair impediment of competitors’ within the meaning of Sec. 20(3a). Rather, the lawmaker identified certain markets as naturally (more) prone to a market tipping.
Therefore, given the important role that the network effects of competitors play to reduce natural tipping tendencies in multi-sided markets, the lawmaker considered it justified to implement a strict liability prohibiting any impediment of the independent attainment of their (own) network effects by competitors.
Criteria for application
Undertaking with superior market power operating on a multi-sided market
The provision is applicable to ‘undertakings with superior market power’ operating in a market within the meaning of Sec. 18(3a) Competition Act (i.e., a multi-sided market, where at least two distinguishable user groups come together, or a network market).
The notion of ‘superior market power’ is tied to the established Sec. 20(3) Competition Act. The wording and systematic anchoring of the provision in Sec. 20 of the Competition Act – whose title reads “Prohibited Conduct of Undertakings with Relative or Superior Market Power” – make it clear that a dominant position within the meaning of Sec. 18, 19 Competition Act or Art. 102 TFEU is not a prerequisite. Rather, according to the lawmakers’ wishes, intervention against the tipping process should be made possible at an early stage, when the undertaking has not yet exceeded the threshold for market dominance.
Within the framework of an overall horizontal assessment, the assessment of whether an undertaking has ‘superior market power’ focuses on a relational comparison and on whether there is a bilateral imbalance of power, considering the relevant turnover within the market and other key performance indicators such as market shares, number of accounts, quantity of user traffic, brand awareness, overall financial strength etc. Akin to the determination of a dominant position, the determining question is whether the undertaking with superior market power enjoys a special position that enables it to behave to an appreciable extent independently of its competitors. Although the opposite has been advocated, it is no prerequisite that the undertaking with superior market power is ‘capable to cause the market to tip’.
Impediment of the independent attainment of network effects by competitors
The conduct at issue is any ‘impediment of the independent attainment of network effects by competitors.’ The wording of the prohibition is rather broad and encompasses any practice which obstructs the independent attainment of their(own) network effects by competitors, including yet unknown practices. According to the legislative materials, the provision specifically targets conduct which (i) hinders multi-homing or (ii) impedes platform migration of users. This focus makes sense, as unimpeded multi-homing and platform migration not only enable the generation of network effects of competitors, but also actively counter any pre-existing (concrete or abstract) tipping tendencies that may exist in the relevant market or evolve in the future. It follows notably that exclusivity agreements and restrictions to portability are prime examples for prohibited conduct under the new Sec. 20(3a) Competition Act.
However, the provision does not entail the use or sharing of third-party network effects and does, therefore, not grant any access or interoperability requests.
Creation of a serious risk of significantly restricting competition on the merits
Being a strict liability tort, the mere creation of a ‘serious risk’ of a ‘significant restriction’ of competition is sufficient to trigger the pre-emptive prohibition contained in Sec. 20(3a) Competition Act. In other words, intervention is desired before actual harm to – often fragile – competition in digital markets occurs, as such damage is often irreversible. By lowering the intervention threshold, conduct that is proven to be dangerous for competition can be prohibited regardless of evidence of concrete effects.
Hence, taking into account the risk of potential over-enforcement (i.e., false positives), the legislator made a conscious decision when faced with the question of the optimal timing of intervention, and clearly opted for early action. According to the lawmaker, shifting intervention to this comparatively early stage is justified by several circumstances that are particular to the relevant markets: (i) the difficulties to predict if, when, and how a market may tip (see above), (ii) the potentially rapid course of a tipping process once it is started, (iii) the often-irreversible effects of tipping, and (iv) the high risk that emanates from the impediment of a competitor’s ability to generate (own) network effects.
Application of the ‘tipping paragraph’ in practice
Within one year of adoption, there have already been three different chambers of the Regional Court of Berlin and, on appeal, the Higher Regional Court of Berlin, deciding on a set of closely related anti-competitive practices and applying the ‘tipping paragraph’ in practice.
The dispute in question concerns a so-called “List First Discount” that the – by far – leading German real-estate platform Immobilien Scout (“ImmoScout”) is offering to its realtor customers. The realtors are granted a 10 percent discount on their monthly advertising fee if they agree to list a specific proportion of their properties for lease and sale exclusively on the ImmoScout platform for a certain initial period.
- Originally, the realtors had to list 95 percent of all their properties for the first seven days of publishing exclusively on the ImmoScout platform.
- After the Berlin Regional Court prohibited that practice by way of an interim injunction on April 8, 2021 (Case 16 O 73/21 Kart), ImmoScout slightly adapted the conditions: Now, realtors had to list 75 percent of their properties exclusively on ImmoScout – for 14 days.
- Unsurprisingly, the same court – but a different chamber – forbade the amended practice as well by an interim injunction dated December 14, 2021 (Case 15 O 290/21 Kart).
- However, ImmoScout continued offering List First Discounts to its realtor customers by introducing an option under which realtors could switch the List First obligations off. As such a mere theoretical possibility to switch the exclusivity obligation off did not change anything meaningful in practice, this led to a third interim injunction by the first chamber of the Regional Court of Berlin, which acted this time in a new composition (Case 16 O 82/22 Kart, judgment of February 24, 2022).
Meanwhile, on February 11, 2022, the appellate court – the Higher Regional Court of Berlin – had already rendered an indicative ruling in ImmoScout’s appeal against the first judgment – announcing its clear intention to reject the appeal by way of a decision without oral hearing as it deemed the appeal “manifestly ill-founded” because the court saw a clear and flagrant infringement of Sec. 20(3a) Competition Act (Case U 4/21 Kart). Following this indicative ruling, ImmoScout withdrew its appeal. Thus, the first judgment on the German tipping paragraph is final. Moreover, ImmoScout withdrew its right to appeal the second and third judgments.
Overall, twelve judges specialized in antitrust law decided in this complex of proceedings, which makes this body of court decisions a highly valuable source for interpreting the new provision. As far as publicly known, these four decisions are the first to apply Sec. 20(3a) Competition Act in practice. All decisions came to the same conclusions, although minor details in the assessment differed. All decisions focused on the possible and likely courses of action of the realtors if they agreed to the ‘List First’ conditions, i.e., the effects of their listing activities on competing platforms. The different chambers agreed that the first days are the most important in the overall listing cycle, and that competing platforms suffer significantly when formerly multi-homing realtors cancel their existing contracts with competitors to exclusively fulfill their demand with ImmoScout (i.e., reduced quantity of multi-homing), or where they receive (i) fewer or (ii) outdated listings (i.e., reduced quality of multi-homing). In particular, in their assessment of whether ImmoScout obstructed its competitors in the generation of positive indirect network effects, the courts recognized the interdependencies between the amount of available (topical) listings and the attractiveness of the platform for searchers of properties (i.e., potential buyers or tenants).
- Reducing the amount or quality of available listings on competing platforms inevitably leads to a negative feedback loop reducing the number of searchers navigating to such competing platforms, which in turn again decreases the platform’s attractiveness for realtors and other advertising customers.
- Conversely, the head start in exclusive, highly topical listings makes the market-leading platform even more attractive for searchers and conversely for advertising customers – creating a vicious cycle.
According to the courts, the result of such effects is the creation of a tipping risk on the multi-sided market for online real estate platforms in Germany.
Judgment of April 8, 2021 (Regional Court of Berlin, Case 16 0 73/21 Kart)
The first judgment on the List First Discount is obviously special because the court had to apply Sec. 20(3a) Competition Act for the first time, with little more than the legislative materials and a few first academic observations as guidance. While the court could theoretically also have resorted to resolving the case under the abuse of dominance provisions (Sec. 19 Competition Act, Art. 102 TFEU) or based on unfair trading law, the court did not shy away from sailing into unexplored waters.
The chamber of three judges ruled that the List First Discount led to de facto exclusivity because property listings generally have little, or at least less, value after seven days. The court could have based this finding on its own experiences in the highly competitive Berlin real estate markets, where attractive listings receive hundreds of leads within only a few hours after insertion on a leading real estate platform. After seven days, a major portion of all listings is offline again, which was not disputed by the defendant. While advocating for a low(er) standard of proof, the chamber did in fact assess the actual foreclosure effects that the plaintiff could already demonstrate since the introduction of the List First Discount scheme (i.e., supranormal decline in customers and revenues). It is noteworthy that the chamber clearly held that there is always a risk of tipping in the area of online platforms. From this, the chamber also derived the urgency and need for interim relief.
Judgment of December 14, 2021 (Regional Court of Berlin, Case 15 O 290/21 Kart)
The chamber that delivered the second judgment could have based much of its analysis on that of the parallel chamber. However, it gave more emphasis to the nature of the provision as a strict liability tort, where the mere abstract capability of creating foreclosure effects must be sufficient to find an infringement. By doing so, the chamber also underlined the parallels between Sec. 20(3a) Competition Act and German unfair trading law (which are also visible in the legislative materials). Moreover, the court emphasized the otherwise existing evidential difficulties before the civil courts, in particular in interim relief proceedings.
The court considered the List First Discount a “prime example” of conduct that the lawmaker intended to prohibit.
Indicative ruling of February 11, 2022 (Higher Regional Court of Berlin, Case U 4/21 Kart)
The third decision, the indicative ruling of the Berlin Higher Regional Court, was the most comprehensive decision (nearly 65 pages). The court clearly rejected the appellant’s (i.e., ImmoScout’s) factual and legal arguments. In fact, the legal arguments were mainly based on criticism that the provision attracted during the legislative procedure – and which the lawmaker rejected by adopting the provision without any changes.
With regard to the material scope of application, the Berlin Higher Regional Court clearly rejected the idea that any concrete tipping tendency is a precondition for applying Sec. 20(3a) Competition Act. Neither the wording nor the systematic interpretation of the provision suggests this, as the provision solely refers to markets within the meaning of Sec. 18(3a) – that is, multi-sided markets or networks – which are, therefore, per se deemed to be (potentially) prone to tipping due to the existence of network effects. Moreover, it would not be in line with the intention of the legislator if measures that pose an inherent risk of tipping were excluded from abuse control from the outset due to the lack of an actually ongoing tipping process. In the same vein, the court rejected the idea that the norm addressee must be in a position to induce the actual tipping process, as any such capability test is reserved to the last stage of assessment – the creation of a ‘serious risk for competition.’
In line with the open wording of Sec. 20(3a) Competition Act, the court further held that the conduct at issue may be any behavior that directly or indirectly makes it more difficult for a competitor to independently achieve network effects. The court rightfully considered positive network effects to be crucial for the success of platform-based business models. In other words, the market must not be influenced in such a way that competing platforms cannot generate a self-reinforcing growth momentum.
However, based on a teleological interpretation, the court considered that only conduct that departs from competition on the merits is covered by the notion of ‘impediment,’ as the Competition Act is not intending to restrict normal competition. Measures that increase the attractiveness of the undertakings own offering while also indirectly restricting competitors in their ability to generate their own network effects are therefore not covered by the prohibition. The control question can be whether the measure to be examined would also make business sense independently of the obstructive effect.
Unsurprisingly, the court confirmed that ImmoScout clearly departed from competition on the merits by employing a targeted exclusionary strategy which “did not aim at generating or preserving own network effects, but is aimed solely at ensuring that competing products are not used, at least temporarily.” As per the court, the goal of the List First Discount is “to worsen the competitive position of the others and thus only indirectly about generating an own advantage.”
As regards the creation of the ‘serious risk for competition’, the court affirmed that it is not required that an actual tipping process has already started. However, the notion of ‘serious risk’ must be interpreted restrictively according to the court, which means that a mere hypothetical possibility of restricting competition is insufficient. The court recognized that courts’ possibilities and insights are limited, in particular in proceedings for interim relief, but emphasized that the lawmaker consciously accepted the associated dangers for any potential over-enforcement. For the purpose of determining the capability of a practice to restrict competition, there needs to be an assessment of the specific circumstances of each case, based on objective and concrete indications, and taking into account the expected effect of the practice and the concrete circumstances of the market. A serious, not merely theoretical, restriction of competition – i.e., not the tipping – must seem already underway. The Higher Regional Court made it clear that even if the practice at issue did not yet translate into concrete losses of existing customers or lost sales, the intervention threshold would already have been reached because the court “perceives the introduction of the discount as a serious attack on multi-homing, which is likely to manifest itself only gradually in further losses”.
In its detailed analysis of whether such risk exists, the court started by stating that a mere predicted emergence of a “simple dominant position” is insufficient for assuming a tipping risk. Rather, the initiation of a self-reinforcing dynamic is required, which may lead to a long-term, entrenched monopoly-like position. The court assumed that positive network effects, economies of scale and scope, and preferential access to data lead to positive feedback loops and, as a result, to a concentration of demand on one platform, which means that remaining competitors or new entrants to the market can no longer acquire customers. The court found that such risk existed in the present case.
Lastly, by way of obiter dictum, the court declared that such an assessment would not differ for the – in this case not contested – other versions of the List First Discount, as they may be even more restrictive due to the extended exclusivity period of 14 days.
Judgment of February 24, 2022 (Regional Court of Berlin, Case 16 O 82/22 Kart)
The fourth decision on the List First Discount – and third interim injunction against it – made clear that the mere hypothetical option for realtors to switch off the List First obligation, together with an option to turn the discount scheme on again at a later point, did not ease the competition concerns. Rather, such purported flexibility, if anything, increased the risks associated with the contract clause: Instead of creating a real opportunity for competitors to target the customers contractually bound to list exclusively on the ImmoScout platform for an initial period, the on-and-off switch likely enhanced economic incentives for formerly undecided realtors to ‘try out’ the List First Discount (to later stick with it). Therefore, the latest version of the List First Discount was capable of even increasing market penetration of the anti-competitive practice.
Moreover, the third judgment also brought into question two interim conclusions of the Higher Regional Court. By sticking to the clear wording of the provision, the judgment advocated a partially lower standard regarding two important points: First, the Regional Court deduced from the legislative materials that the hindering of multi-homing, as well as the impediment of platform switching, are forms of conduct which have been identified as inherently “problematic,” and thus are not competition on the merits by their very nature. As a consequence, in the case of such “unwritten rule examples,” there is no need to perform the further competition on the merits analysis that the Higher Regional Court deemed necessary by resorting to a teleological interpretation of the provision (see above). In addition, the chamber not only found the first unwritten rule example – hindering of multi-homing – to be fulfilled, but also held that ImmoScout’s practice impeded platform switching, Second, the judgment clarifies that while the provision aims at preventing a monopoly as the ‘end state’ of a tipped market, it does so by prohibiting the weakening of the identified markets’ structure. Thus, there is no need to assess whether such an end state will likely be reached in the (near) future, or whether the tipping process has already started, or whether a self-reinforcing dynamic has been initiated. This is in line with the chamber’s finding that the legislator deemed multi-sided platform markets per se prone to a tipping risk. Again, this is a lower standard than that advocated by the Higher Regional Court in its interim conclusions on the case.
First practical experiences show that Sec. 20(3a) Competition Act is a powerful tool to tackle competition concerns in multi-sided markets below the threshold of the traditional abuse of dominance prohibitions. Its clear focus on multi-homing hindering practices in markets that are deemed to be fragile and prone to concentration makes it an efficient instrument, notably in proceedings for interim injunctions. Conversely, the new provision places considerable compliance requirements on market-leading platforms.
Threshold and burden of proof
Due to the link to a position of (merely) ‘superior market power’ in a market that is deemed, by law, as abstractly prone to tipping, its nature as a strict liability tort – meaning, that concrete effects or actual restrictions need not be proven – and the open wording as regards the restrictive practices, Sec. 20(3a) contains an intervention threshold lowered in several ways. The prohibition provides for swift, pre-emptive intervention before any harm to competition occurs. In particular, there is no need for a sudden shift of market shares with regard to performance indicators.
This comparatively low intervention threshold is precisely why the provision has been harshly criticized by academic voices. However, the first decisions dealing with such criticism clearly reject it. In the view of the authors, the courts’ rejection of such criticism is well-founded, as any limiting interpretations that have been advocated prior to the adoption of Sec. 20(3a) Competition Act (i) can naturally not be transferred without scrutiny to the post-adoption period and, moreover, and (ii) cannot be reconciled with the unambiguous intention of the lawmaker, the clear wording of the provision, and its systematic and teleological interpretation. It can only be assumed that courts will follow the approach taken by the twelve judges of the Berlin courts in the future as well.
In addition, the first judgments indicate that economic expert opinions will not necessarily be needed to convince a court of an infringement of Sec. 20(3a) Competition Act. In fact, in the course of the first proceeding (Case 16 O 73/21 Kart), none of the parties were advised by competition economists. This is even more true if the tendency of applying fair trading law principles continues.
Suitability for private enforcement and interim relief
Some commentators argue that the new provision is only suited for public enforcement by the Federal Cartel Office because it requires lengthy in-depth market analyses. The opposite actually is true. The comparatively low intervention threshold and burden to bring private actions, along with its goal to prevent the occurrence of actual effects on competition, make Sec. 20(3a) particularly suitable for interim relief proceedings. In fact, in view of the intended goal of the provision, its activation in private interim relief proceedings or in interim measures imposed by the Federal Cartel Office (Sec. 32a Competition Act) may well become the primary scope of application of the provision. The procedural urgency needed for interim relief actions or interim measures is inherent to the material scope of the provision.
Need for rigorous compliance
The effective new instrument, the low intervention threshold, and the ability to apply it in a cost-efficient manner in private actions by competitors requires particular attention from market-leading platforms in multi-sided markets or leading networks.
A number of sectors can be identified in which the provision may become relevant, for example, online marketplaces and comparison sites, platforms of all genres (dating, ride-sharing, transport, food delivery), career networks and online recruitment platforms, messenger services, B2B distribution platforms or networks, online search services, advertising networks and ad tech services, and so on and so forth.
Market-leading operators should carefully analyze whether practices that restrict free multi-homing or switching of platforms are compliant with the new provision. Practices that may still have escaped one of the established case groups of the abuse of dominance provisions, or that require in-depth economic analysis (e.g., exclusivity rebates with a low market coverage), may now be prohibited more easily and expose undertakings to the risks of incurring penalties, litigation, and reputational damage.
*Dr Ann-Christin Richter is a Partner and Philipp Westerhoff is a Senior Associate in the Berlin office. Transparency: The authors represented the German real estate platform Immowelt in the proceedings that led to the four decisions discussed in this article.
 Federal Law Gazette 2021 I No. 1 of January 18, 2021, page 2 et seq. The new provision reads as follows: “An unfair impediment within the meaning of subsection (3) sentence 1 shall also be deemed to exist where an undertaking with superior market power on a market within the meaning of Section 18(3a) impedes the independent attainment of network effects by competitors and in this way creates a serious risk of significantly restricting competition on the merits.” For an English translation of the Competition Act, see https://bit.ly/3oZ5VaT.
 Government reasoning of the 9th amendment of the Competition Act of November 7, 2016 (BT-Drs. 18/10207), page 49
 Government reasoning of the 9th amendment of the Competition Act (fn. 2), page 50.
 See Schweitzer/Haucap/Kerber/Welker, Modernising the law on abuse of market power (“Modernisierung der Missbrauchsaufsicht für marktmächtige Unternehmen”), August 29, 2018, page 55 (available online: https://bit.ly/3gRLqZ8).
 See Schweitzer/Haucap/Kerber/Welker (fn. 4), page 60.
 See id., (fn. 4), page 59.
 In addition to Sec. 20(3a), the legislator introduced notably a third category of market power that is subject to additional obligations – so-called ‘undertakings with paramount significance for competition across markets’ (Sec. 19a). See Höppner, Digital upgrade of German Antitrust Law - blueprint for regulating systemic platforms in Europe and beyond?, March 9, 2020, for an overview on the overall amendments (available online: https://bit.ly/33tV6Gh).
 Government reasoning of the 10th amendment of the Competition Act of October 10, 2020 (BT-Drs. 19/23492), page 82.
 See Schweitzer/Haucap/Kerber/Welker (fn. 4), page 62. However, the wording of the study is rather abstract and refers to the existence of such risks in case of “markets, which are characterized by strong positive network effects”. See also Esser, Kartellrecht und Tipping – Kippt die Missbrauchsaufsicht mit Einführung des § 29 Abs. 3a GWB-RefE in eine präventive Regulierung?, Commemorative Publication for Wiedemann, Das Unternehmen in der Wettbewerbsordnung, 2020, page 296.
 That is, multi-sided markets or networks within the meaning of Sec. 18(3a) Competition Act.
 However, unlike Sec. 20(3), Sec. 20(3a) is applicable to all undertakings with superior market power, not only undertakings with superior market power ‘in relation to small and medium-sized competitors.’
 Government reasoning of the 10th amendment of the Competition Act (fn. 8), page 82.
 See the references to ImmoScout’s legal expert opinion in Higher Regional Court of Berlin, Indicative Ruling of February 11, 2022, Case U 4/21 Kart (available online: https://bit.ly/3wJoLYc), at para. 54.
 See id., (fn. 13), at para. 45, and 54 et seq.
 Government reasoning of the 10th amendment of the Competition Act (fn. 8), page 82, referring to Crémer/de Montjoye/Schweitzer, Competition policy for the digital era, page 57 et seq. (available online: https://bit.ly/3JDB23x).
 Government reasoning of the 10th amendment of the Competition Act (fn. 8), page 82.
 Id., (fn. 8), page 83; see also Higher Regional Court of Berlin, Indicative Ruling of February 11, 2022, Case U 4/21 Kart (fn. 13), at para. 159.
 Government reasoning of the 10th amendment of the Competition Act (fn. 8), page 83.
 In one of the two antitrust chambers of the Berlin Regional Court, new members were appointed in the course of the proceedings. Therefore, an entire new chamber composition decided on the third injunction.
 Higher Regional Court of Berlin, Indicative Ruling of February 11, 2022, Case U 4/21 Kart (fn. 13), at para. 154.
 Id. at para. 154.
 Id. at para 174.
 Id. at para. 217.
 Esser (fn. 9), page 301, argues that it would be hard even for the Federal Cartel Office with all its investigatory powers to conclude proceedings based on Sec. 20(3a) Competition Act.