• No results found

Operationalization - Algorithmic Exploitation under Art.102 TFEU

4. An Ordoliberal Doctrine of Algorithmic Exploitation

4.4. Operationalization - Algorithmic Exploitation under Art.102 TFEU

In this section, I will explore how the findings presented above can be embedded into a theory of harm within the current competition law framework.

This theory of harm is based on Art.102(2)(a) TFEU, which, inter alia, prohibits the imposition of “unfair selling prices” by one or more dominant undertakings.

4.4.1. ESTABLISHMENT OF COLLECTIVE DOMINANCE

As demonstrated above, the finding of a collective dominant position within the scope of Art.102 TFEU overlaps with the economic conditions necessary for the non-cooperative collusion to occur. This burden of assessment is thus not discussed in further detail in this thesis.

A point worth mentioning is that increasing data power may lead to higher entry barriers and could drive non-equipped competitors out of the market eventually, leading to higher market power.

4.4.2. QUANTITATIVE ASSESSMENT OF “UNFAIR PRICES

Regarding the interpretation of when a price is “unfair”, the ECJ has held that a price may be unfair if it is excessive in relation to the economic value of the product supplied.152 This excessiveness can be determined by an economic assessment of the product’s profit margin.153

151 This, however, may be beneficial for consumers and thus justified. See in this regard section 4.5 on efficiency considerations. However, it must be noted that, if the market price itself does not reflect competitiveness (as it is the case in tacit collusion scenarios), the secret deal only constitutes a discount from a supra-competitive equilibrium. Deadweight losses may be reduced compared to a purchase for the market price, but the discount price may still be supra-competitive itself. These hybrid situations of tacit collusion combined with personalized pricing would then need to be carefully assessed.

152 Case 27/76, United Brands, para. 250.

153 Ibid, para. 251.

When excessiveness is affirmed, the unfairness may lie in the price itself or can be determined by comparison to the prices of comparable products.154

Although guided by a calculation of excessiveness, this concept appears to be vague and unspecific. Besides this problem of legal uncertainty, there are two more challenges when applying this quantitative approach.

First, in digital markets this concept can lead to practical problems. In a digital environment, the competitive counterfactual is very difficult to establish due to specific factors like network effects and economies of scale and scope.155 There may also be no comparable products or services and the economic evaluation of a product or service may also not be possible. For instance, assigning an economic value to data is highly complicated.

Second, from an ordoliberal understanding of competition as a dynamic process, price controls are generally undesirable. The ordoliberal state has the role to set up and maintain the competitive order, not to intervene in the market.156 The establishment of a competitive counterfactual necessarily includes a presumption of knowledge which, according to the ordoliberal understanding, can only derive from the competitive process itself. 157 As described above, the outcome of a competitive process is unpredictable and thus unquantifiable from an ex-ante standpoint. Hence, as an ex-post intervention to correct market results, price controls should be regarded ultima ratio.158

Consequently, the German ordoliberals agreed to the inclusion of the prohibition of exploitative practices in the Rome Treaty only because it was confined to particularly excessive conduct.159

154 Ibid, para. 252.

155 Gal/Petit (2021), 9, “when competition is for the market, the counterfactual is a monopoly structure”; For competitive aspects in digital markets see Digital Competition Expert Panel (2019).

156 Bonefeld (2012), 634; Anchustegui (2015), 143.

157 The “as if” standard to establish a competitive counterfactual was originally proposed by Eucken for dealing with natural monopolies, but has later been abandoned, see Behrens (2018), 18; Anchustegui (2015), 146, emphasizes that suppressing price controls facilitates trust in the market-price system.

158 Fuchs (2019), para. 172.

159 Behrens (2018), 12, 25, mentions that the French delegation opted for even wider price controls.

4.4.3. PROPOSAL FOR A QUALITATIVE ASSESSMENT

To avoid the mentioned challenges, I propose a new benchmark for the unfairness of prices.

The current quantitative approach appears to be too narrow. Thus, the interpretation of

“unfairness” must be scrutinized and replaced by a broader and more flexible benchmark.

A similar problem was subject of a case under German law regarding an alleged abuse of dominance by Facebook, where the company was accused of imposing inappropriate trading conditions on consumers. The German Federal Cartel Office (Bundeskartellamt) had based the finding of inappropriateness on normative values of the GDPR, whereupon the Higher Regional Court Düsseldorf (OLG Düsseldorf) in second instance criticized the lack of investigations on the competitive counterfactual.160 The German Federal Court of Justice (Bundesgerichtshof) then clarified that a concept comparing the actual conditions with those under effective competition is not suitable if a specific determination of these competitive conditions is hardly possible.161

This concept, in my opinion, is transferable to the subject of this thesis. Although the Facebook case concerned § 19 GWB, the German equivalent to Art.102 TFEU, the latter must also be accessible to a qualitative concept in areas where the quantitative concept cannot provide adequate results, for instance in digital markets.

As shown, it is precisely these digital markets that are most vulnerable to algorithmic tacit collusion. Setting price-benchmarks based on ex-post calculations then bear a high risk of enforcement errors due to the aforementioned characteristics.

I thus propose to use the restriction of the consumer’s sovereignty as a qualitative yardstick and benchmark for the unfairness of prices. This would link the undesirable correction of market results to the deliberate disturbance of the ordoliberal construct of dynamic competition. It would focus on the procedural means of price setting rather than on the mere economic effects.

When using this benchmark, prices imposed by dominant undertakings on consumers by means

160 OLG Düsseldorf, 26.08.2019, VI-Kart 1/19 (V), EuZW 2019, 779 para. 20 – Facebook.

161 BGH, 23.06.2020, KVR 69/19, para. 82, the concept of so-called “as-if-competition”.

of algorithms, while eliminating price competition and restricting consumer sovereignty, must be regarded as unfair and thus abusive.

Overall, this would revive price controls as a competition law instrument while at the same time respecting ordoliberal key concepts of consumer sovereignty and process-oriented competition law assessment. However, there are challenges for practical enforcement which will be explored in section 5.