Immaterialgüter- und Wettbewerbsrecht

Algorithmic Personalized Price Discrimination in the Data Economy. Rethinking the EU Competition Policy vis-à-vis New Exploitative Threats

Algorithmic personalized pricing will be common in the digital era, leading to efficiencies but also to novel forms of discrimination with exploitative effects. This research aims to answer whether the EU laws of competition and unfair commercial practices are suitable to deal with this new reality.

Last Update: 27.04.21

The digitization of society has increased the amount of data extracted from individuals participating in the digital world. This has led to the reduction of data processing costs through Machine Learning, improving the accuracy of consumer profiling in order to better predict their behaviour and willingness to pay. In addition, this has allowed the automation of tasks previously performed by humans, such as deciding the price consumers are going to be charged for a product. This form of pricing is identified here as algorithmic personalized price discrimination (“APPD”), which can be understood as an extremely atomized form of third-degree price discrimination, with ambiguous effects.

While APPD may improve allocative efficiencies and be part of procompetitive strategies (e.g. to attract new consumers and thereby gaining sufficient network effects to challenge an incumbent), it may also lead to harmful effects. First, the extraction of consumer surplus through APPD may be the result of less consumer autonomy. This would be the case if APPD is possible due to the opacity of algorithmic prices (i.e. information failure) and/or economic dependence of market players (i.e. lock-in). In these scenarios, the underlying problem of discrimination —rather than other exploitation forms— is whether a differentiated price scheme may create disadvantages to consumers vis-à-vis each other, ultimately distorting the competition process and leading to regressive distributive effects. In the data economy, where multi-sided platforms could be able to shape the marketplace, this scenario may become the rule.

Art. 102 TFEU seems adequate to tackle only few of the possible harmful hypotheses caused by APPD, for which it would be required an estimation about when the disadvantages generated by discrimination can impact the competitive process and consumer choice. Yet, defining this yardstick might be an extremely difficult task in the context of digital markets. On the other hand, the EU law on unfair commercial practices (UCP) emerges as a possible way to tackle some of the opacity-based issues, as it empowers consumers and set transparency standards about prices. However, not every UCP infringement committed by a dominant firm should be taken as exploitative, as this would create a conflict of laws and reduce dynamic competition. A balance is needed.

Accordingly, two policy recommendations are presented. On the one hand, the research favours a regulatory ban of certain elements of APPD formation in order to avoid opacity. On the other hand, as exploitative APPD might be possible without dominant position as traditionally understood, the rule of dominance established in Art. 102 TFEU should be reinterpreted. An exploitation should be possible to be found when there is dependence and the competitive process is being impaired. This can coincide with an UCP infringement or not; and could take place in spite of allocative efficiencies.


Doctoral Student

Germán Johannsen

Main Areas of Research

II.3 Vernetzte Datenwirtschaft