In new research, Wolfgang Kerber argues that the European Union’s decision to require a third option in Meta’s “consent or pay” model focuses too heavily on user consent and ignores the broader goal in the Digital Markets Act of reducing gatekeeper advantages and increasing competition.
The European Union Digital Markets Act’s ex-ante regulation of the behavior of gatekeeper platforms is often seen as the most ambitious approach to challenging the power of large tech firms. But everything depends on the effectiveness of its enforcement.
For example, Article 5(2) of the DMA intends to limit the data power of gatekeepers and achieve more contestability and competition as well as more choice for the consumers regarding how their personal data is collected and used. It wants to ensure that gatekeepers cannot use their troves of data to erect entry barriers and foreclose other competitors.
In one of its first decisions about the enforcement of the DMA, the European Commission decided in April 2025 that Meta’s simple “consent or pay” model did not comply with Article 5(2). There is a broad debate among scholars and policymakers whether it would be sufficient to offer consumers only a choice between (a) a service with consent to use their personal data for behavioral advertising (Meta: “With Ads” option) and (b) an equivalent service where the user pays Meta a fee not to use their personal data for advertising (Meta: “Subscription for No Ads”).
The European Court of Justice (ECJ) in its well-known “Meta” decision in July 2023 decided that “consent or pay” might be compatible with European data protection law, the General Data Protection Regulation (GDPR), if such a fee is deemed “necessary” and “appropriate.” However, the Commission chose to reject this approach, and instead claimed the need for an additional, third option. This option would offer a less personalized advertising model (e.g. contextual advertising) for free. This reflects a solution strongly recommended by the European Data Protection Board (EDPB) for “consent or pay” models on all large online platforms to ensure users’ valid consent. Indeed, Meta has switched to a consent or pay model with such an additional third option without behavioral advertising (“Additional Ads”). However, Meta rejects the position of the Commission and the EDPB as not being compatible with European law, with pending litigation before European courts against the decision of the Commission.
In a new paper, I consider whether or not this additional third option is a sufficient remedy for limiting gatekeepers. While the Commission’s decision presents potential advantages that reduce data accumulation and advertising, ultimately, most users won’t opt for the alternative instead of the familiar “With Ads” option. This decision reflects the Commission’s prioritization of data protection over contestability.
The rationale for Article 5(2)
What are the real issues in this case? What is the link to the power of large tech firms, and how and why does data protection law play such a big role in this decision about compliance of Article 5(2) of the DMA?
The creation of Article 5(2) drew from the 2019 German case against Facebook. Therein, the German competition authority (Bundeskartellamt) decided that Facebook’s combination of personal data from different sources without an additional choice for users is an exploitative abusive behavior of a dominant firm in competition law, because this lack of choice is a violation of data protection law. This led to Article 5(2)’s prohibition on gatekeepers engaging in various types of data combinations, especially for behavioral advertising, “unless the end user has been presented with the specific choice and has given consent,” according to EU data protection law.
However, the main rationale of Article 5(2) is based on the problem that gatekeepers have large data advantages (also via data combination) compared to other firms. This leads to exclusionary effects, and, in particular, reduces its contestability. For example, Meta can use the vast troves of data it collects on its social media platforms to feed users ads or products from Facebook Marketplace. This gives Facebook Marketplace and Meta an advantage over other markets for second-hand goods like eBay. Therefore, Meta’s data power has far-reaching negative effects on competition (both on advertising markets and other markets), but it can also have direct negative effects on users. The wealth of data in consumer profiles can be used for informational and behavioral manipulation to steer users through behavioral advertising to products that are not in their best interests but rather those of advertisers.
This raises many additional concerns about exploitative behavior from a consumer protection and privacy perspective, and the current debate on artificial intelligence agents reinforces these concerns. Therefore, limiting data power, especially through less data combination, might be very important, and would fit entirely the objectives of the DMA of more contestability and fairness. But what did the Commission do in its Meta decision?
Weighing data protection against contestability
Most importantly, the Commission focused entirely on the choice of data combination for users and not on the objective of contestability. The Commission interpreted “specific choice” and “consent” (under the GDPR) in Article 5(2) as two independent conditions. Its reasoning for dismissing Meta’s “consent or pay” model only used the logic of data protection. When assessing the condition of “consent,” the Commission applied typical data protection law standards, such as whether there is an imbalance of power between the users and the platform that allows the platform to set unfair terms for the use of data. In its assessment of “specific choice,” the Commission argued that any monetary payment would make the payment option without Ads not equivalent to the free With Ads option. This leads to a violation of the requirement in Article 5(2) that Meta has to offer an equivalent less-personalised option. Such a far-reaching formalistic interpretation of the condition “specific choice” by the Commission leaves de facto no room for any monetary payment in Article 5(2), and therefore leads to a direct rejection of “consent or pay” models. This also implies that the Commission did not assess whether the fee in the payment option is “appropriate,” as suggested by the ECJ.
The problem is that unless users are voluntarily choosing the options that reject data collection to be used for behavioral advertising, Meta’s power hasn’t really been limited. Only if many or most users choose the payment option or the free option with less personalized advertising can the Commission’s decision significantly reduce the data power of Meta, leading to more contestability. However, in its decision, the Commission does not address this problem and how to increase the number of users opting out of data collection. For example, it does not consider increasing the number of users that reject the “With Ads” option by forcing Meta to lower the monetary price for the payment option. In contrast, the Information Commissioner’s Office (ICO), which serves as a data protection authority in the United Kingdom, has accepted the simple “consent or pay” model of Meta but required a much lower price than in the EU.
Of course, the Commission’s strategy of insisting that Meta offers an additional free option without behavioral advertising can be seen as a step for inducing more users to reject the option with behavioral advertising. However, it is unclear whether this is enough for breaking the habit of many users to choose the familiar With Ads option of Meta, and, of course, there will be users who prefer a more personalized advertising. The problem, however, is also that in Meta’s current “consent or pay” model, the design of the choice architecture—how users are presented with the options—is still biased against this additional free option with less personal data. Meta still nudges its users to its With Ads option and makes it difficult to find and choose the free third option without behavioral advertising. The Commission, so far, has not tried to enforce sufficiently the anti-circumvention rules of the DMA against behavioral manipulation (Article 13 DMA). Therefore, it is unclear whether its approach will even effectively protect data privacy.
Conceptually more important, however, is that with its insistence in Article 5(2) on following the suggested solution of the EDPB, the Commission has de facto assumed only the role of an additional enforcer of data protection law. It is surprising that in its assessments of “specific choice” and “consent,” the DMA objectives of “contestability” and “fairness” do not play any explicit role. Although the Commission emphasizes that the first condition “specific choice” is DMA-specific and does not relate to the GDPR, it does not present an approach that addresses how the contestability objective can be achieved. In that respect, the Commission missed an important opportunity to clarify the role and approach of Article 5(2).
Targeting contestability of gatekeepers
Therefore a more targeted implementation approach that does not only rely on data protection law is necessary for reducing the data advantages of gatekeepers. The Commission should focus more on the direct prohibitions of data combinations in Article 5(2) as the basic approach, and then develop criteria that establish whether, to what extent, and how the gatekeepers should be allowed to use the exception of getting consent for the combination of data. Since data power and data accumulation can lead to large dangers for contestability and data protection, it can be argued that this exception should be applied in a much more narrow way or even eliminated under certain conditions.
Would such an approach to achieve more contestability lead to a too large limitation of the free choice of users? Do we have a conflict between both objectives, and therefore need some balancing between them? First, it is not clear whether deeper consumer profiles through more data combination will lead to better consumer decisions (via better matching of products and preferences) or, vice versa, to more harm to consumers due to a more effective manipulation of their decisions. Therefore, less data combination might also have positive effects on consumers, also with regard to data privacy, and from more competition. Therefore, Article 5(2) DMA might not lead to a trade off problem between competition policy and data protection. Instead, positive synergy effects might prevail. But even if this is not so clear, it can still be argued that the risks of gatekeepers’ large data power makes achieving more contestability more important than an unlimited free choice of individual users in all cases.
Such an approach could be implemented in different ways. One possibility is to establish the free option without behavioral advertising as a default solution. This could imply that the Commission develops a much more active approach regarding the design of the choice architecture, such as including information about the advantages and disadvantages of different options, as well as warnings about the risks of giving gatekeepers too much data power. Based also upon insights from empirical behavioral research and public discussions, this might lead to the formation of a broad opinion and a user habit of choosing this free, less-personalized option as a de facto default solution. It might also be possible to limit more directly the possibilities of gatekeepers to get consent for data combination through additional conditions.
To what extent such an approach is legally possible with the current version of Article 5(2) or requires a revision of this obligation through further legislation has to be discussed by legal scholars. However, such an approach for more contestability would ensure that Article 5(2) and the Commission as its enforcer can play an independent effective role in the interplay of competition policy and data protection law for dealing with the data power of large tech firms.
Author’s Disclosure: The author reports no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
