Fairness in the Digital Markets Act

Keywords: Digital Markets Act – platform economy – fairness – equity – distribution – legal standard.

The notion of “fairness” is ubiquitous in the Digital Markets Act (“DMA”).[1] Yet, its meaning remains elusive. This piece considers different ways to interpret fairness in the context of the DMA and explores whether the notion could become an evaluative principle for gatekeeper compliance with the obligations in arts 5 and 6.

The root word “fair” appears 90 times in the DMA, in remarkably diverse contexts.[2] The legislator tackles “unfair practices” that could “undermine the contestability of markets”,[3] and situations in which “market processes are […] incapable of ensuring fair economic outcomes”,[4] but also contractual restrictions that “unfairly limit the freedom of business users”.[5] Vertically integrated gatekeeper platforms are prevented from “unfairly benefitting from their dual role”[6] and all gatekeepers  are prevented from “achieving, by unfair means, an entrenched and durable position in [their] operations”.[7] In the context of market investigations, the European Commission is required to find out, among other elements, “the degree to which prices are fair and competitive”.[8] Conditions and parameters for product and content rankings for core platform services of gatekeepers should be “generally fair and transparent”.[9] Also, gatekeepers are required to “provide access, on fair, reasonable and non-discriminatory terms” to certain ranking, query, click and view data to other undertakings providing similar services.[10]

The fairness terminology is peppered all over the text. The contexts where it lands are so diverse that doubt arises as to whether it is used as a mere condiment to give flavour to more concrete objectives like contestable markets and interoperability. But fairness might also be the subtle star ingredient of the ensemble, providing substance and coherence to the rest.[11] While the notion of fairness certainly has an emotional appeal – who could be against a fair economy? – it is unclear whether it holds a meaning concrete enough to produce predictable outcomes when litigation ensues regarding the art. 5 and 6 obligations of the DMA.

In general European competition law, fairness has been interpreted as a reference to broad guarantees of equal economic opportunities and equal procedural treatment.[12] When it comes to substantive questions of law, fairness has been depicted as too indeterminate to be used as an evaluative principle, in particular because there is no agreement on its concrete meaning.[13] In the context of the DMA, the fairness terminology could be dismissed as a mere instrument to bring political consensus during the extraordinarily fast adoption of the text.[14] And yet, the references to fairness are so ubiquitous in the DMA that one could also be tempted to use them to delineate substantive gatekeeper obligations.

The DMA suggests a somewhat positive definition of fairness. The concept is differentiated from the equally pervasive but less controversial notion of contestable markets, as the new obligations for gatekeepers “correspond to those practices that are considered as undermining contestability or as being unfair, or both”.[15] Although fairness and contestability are not the same, they are intertwined: “The lack of, or weak, contestability for a certain service can enable a gatekeeper to engage in unfair practices. Similarly, unfair practices by a gatekeeper can reduce the possibility of business users or others to contest the gatekeeper’s position”.[16]

The legislator comes close to a proper definition, but ultimately fails to pierce the veil of vagueness around the notion:

“For the purpose of this Regulation, unfairness should relate to an imbalance between the rights and obligations of business users where the gatekeeper obtains a disproportionate advantage. Market participants […] should have the ability to adequately capture the benefits resulting from their innovative or other efforts. Due to their gateway position and superior bargaining power, it is possible that gatekeepers engage in behaviour that does not allow others to capture fully the benefits of their own contributions, and unilaterally set unbalanced conditions for the use of their core platform services or services provided together with, or in support of, their core platform services”.[17]

How to ascertain such an imbalance? The text provides some form of legal test in the context of software application (i.e. “app”) stores. According to the DMA, an “imbalance in bargaining power” between gatekeepers and business users would facilitate unfair contractual conditions for the latter. It mentions “benchmarks” that “can serve as a yardstick to determine the fairness of general access conditions”.[18] These are:

“prices charged or conditions imposed for the same or similar services by other providers of software application stores; prices charged or conditions imposed by the provider of the software application store for different related or similar services or to different types of end users; prices charged or conditions imposed by the provider of the software application store for the same service in different geographic regions; prices charged or conditions imposed by the provider of the software application store for the same service the gatekeeper provides to itself”.[19]

And yet, none of these comparisons provides a clear benchmark for fair prices or conditions. Finding an “imbalance between rights and obligations” leaves a wide margin of appreciation for any case handler or, in the case of an appeal, any judge tasked to determine the scope of gatekeeper obligations. 

Another avenue for interpreting fairness is to examine its translation in the different EU languages. While the German version of the DMA uses the same word fairness as the English text, the French, Italian, and Spanish texts offer a more interesting linguistic choice. Fair is translated here as équitable/equo/equitativo. But in the French language, the word fair can also be translated in other legal contexts as juste, loyal, and even raisonnable.[20] That means that the French language has several partial translations of the broader Anglo-Saxon concept of fairness, and the legislator has chosen équitable among the various contenders. Fairness in the DMA is about equity.

Equity is not equality. While under formal equality, individual attributes of subjects are purposefully ignored for the distribution of legal entitlements (e.g., religion and voting rights), under equity, entitlements are distributed with particular attention to an attribute of the subject (e.g., student rebates).[21] This is exactly what the DMA does. It starts with a differentiation between gatekeepers and other market participants, and then distributes entitlements accordingly. The differentiation is based on an unequal distribution of market power and economic rent between the market participants. Gatekeepers have accumulated more market power than non-gatekeepers, and this discrepancy is exacerbated in the context of network externalities, increasing returns, and user biases.[22] According to the DMA, the difference in market power could lead to an entrenchment of the powerful position of the gatekeepers and to disadvantageous trading conditions for the non-gatekeepers.[23] Therefore, the gatekeepers must comply with new obligations, some of which directly translate into new entitlements for certain non-gatekeepers. An example for this could be the mandatory data sharing for gatekeeper search engine data, where smaller competitors of a search engine receive the entitlement to access useful data that they did not create on their own.

Fairness in the DMA means that firms receive different entitlements according to their market power. Those with more market power receive less entitlements and see their economic freedom restrained, and those with less market power receive more entitlements. But although the general direction of the entitlement distribution in the DMA becomes clear, the notions of fairness and equity as such still do not indicate any optimal distribution of entitlements. We only know that those with more market power are likely to receive less entitlements, similarly to the “special responsibility” for dominant undertakings under art. 102 TFEU, or the obligations under the essential facilities doctrine.

The indeterminacy of fairness as an evaluative principle is revealed when comparing it to another principle in competition law, the consumer welfare standard. To simplify, suppose that what is distributed by the DMA is not complex obligations and entitlements, but plain economic rent.[24] From that perspective, the obligations and entitlements of the DMA are a tool to reach a different rent allocation. Now, the consumer welfare standard is about increasing the rent of a predefined group within society, consumers.[25] If gatekeeper conduct were evaluated under a consumer welfare standard, then anything that is known to increase the rent of consumers (as a group, since individuals are averaged away) should be allowed. This is a benchmark that generally makes it possible to distinguish desired from undesired firm conduct, as every instance of conduct is either indifferent to consumer welfare, increases it, or diminishes it. Of course, this is an oversimplification; it is debatable which aspects of human flourishing are part of consumer welfare, and to what extent one should care about today’s consumers compared to future generations. But the baseline is that, under the consumer welfare standard, there is something to be maximized, and most often it is plain, commensurable rent for consumers.

The problem of the fairness notion presented in the DMA is that there is nothing to maximize. The fair distribution of rent is a priori unknown. Although it is quite clear between whom the distribution takes place – between gatekeepers and potential entrants, between gatekeepers and business users, and between gatekeepers and private users –the DMA simply offers no commensurable variable that indicates when fairness in rent distribution is reached.[26] Therefore, under the fairness notion as presented in the DMA, the optimal rent allocation between trading partners in the platform economy remains enigmatic, a bit like the exact amount of the “fair share” for consumers in art. 101(3) TFEU[27]. The only hint is that “Market participants […] should have the ability to adequately capture the benefits resulting from their innovative or other efforts”.[28]

With that in mind, it should be added that a predefined distribution of economic rent could easily become a test for compliance with a policy goal. For example, the social policy of a government can be evaluated according to the reduction of people living under a predefined poverty threshold, or according to the evolution of the Gini coefficient of the country. These are commensurable items that indicate whether a desired distribution of income between selected agents has been reached or not. But the DMA does not designate such a target distribution of rent between trading partners. And in any case, would that be desirable?

The issue with competition law is that pre-defined rent allocations between trading partners might invalidate the very essence of what is protected. The rent allocation that competition law strives for is ever evolving, because the environment that induces consumers to formulate their demand, and the environment that allows producers to propose an offer, constantly evolve too. Protecting competition means making it possible that firms gain market power and economic rent when they are good at aligning their offer to the demand. Under competition law, the merit of a firm to receive rent is ever changing. A fixed rent allocation is incompatible with that essential premise.

But the DMA is not competition law. In the context of increasing returns and the network externalities of the platform economy, market power and economic rent end up in the hands of some, not only because their offer best aligns with demand, but also because of the winner-take-all characteristics of the industry. In the perspective of the DMA, this renders the current rent distribution between trading partners too static, and therefore undesirable. It would be more dynamic if it were a pure product of merit and less a product of industry characteristics. But at the same time, the DMA does not indicate which distribution would align with merit. That is justified because merit is constantly evolving.

This leads to an interesting distinction between ex ante and ex post competition policy. In the DMA, an unequal distribution of market power and economic rent is the distinguishing criterion that leads to different entitlement regimes. Meanwhile, general competition law enables firms to gain market power and economic rent if their offer aligns well with demand. While the DMA distributes entitlements according to market power, competition law allows market power to be allocated according to the purchasing signals of consumers. In that context, the notion of fairness fails to separate the subjective feeling of an imbalance in rights and obligations from anything more tangible.

In the end, even the most sophisticated legal tests are designed to crystallize reality into a binary world in which conduct can be qualified as legal or illegal. To generate such a test for an equity-oriented notion of fairness, it is essential to know (1) between which agents the distribution ought to take place, (2) what is distributed and how to measure it, and (3) which distribution can be considered fair. The DMA is quite clear about the first point, as it defines gatekeepers and non-gatekeepers. There is less clarity about the second point, but it seems that the object of distribution is economic rent. However, there are no details on the third point, and even outside of the DMA, there is no canonical understanding of a fair distribution of economic rent. Therefore, fairness as presented in the DMA is not a legal standard, but a gut feeling.

The only gut feeling with authority is that of a judge. That is why the fairness language of the DMA ultimately confers power from economists to lawyers in competition agencies and courts, who will be tasked to find a “significant disequilibrium in rights and obligations”. They might use fairness as an interpretative vehicle to bring in their own allocative taste, but they might as well thoroughly ignore the notion.

Interestingly, in other branches of law, qualitative and equity-based judgements are rendered all the time.[29] In competition law, one can only wonder what makes us cherish so much the idea that legality must be calculable, that there is something unidimensional to be maximized, and that there is a single social optimum to be reached. In that sense, my above argumentation subscribes to a rather archaic legal worldview, whose essential raison d’être seems to be its ease of administration.