Behrang Kianzad, researcher at Lund University School of Economics and Management, ponders the idea of fairness in EU regulation. It will be difficult to find someone who would argue in favour of a world that is unfair, yet the idea of fairness is not as universal as it may seem, nor is it easy to operationalise. The Commission attempts to do so in new regulation such as the DMA. Important questions remain: what is fair, and what limiting principles are there? Hereunder, Behrang shares his thoughts about the concept of fairness, how it found its way into regulation, and how we can operationalise the concept in a consistent manner.
By Behrang Kianzad
Followers of SCiDA blog on the Digital Markets Act (DMA) are probably aware that DMA pursues distinct objectives: fairness and contestability. Recital 33 argues that for the purpose of the DMA, unfairness should relate to imbalance between the rights and obligations of business users where the gatekeeper obtains a disproportionate advantage. It is not important whether the gatekeeper derives this advantage from anti-competitive behaviour or if it is related to the characteristics of the market. The existence of an undue competitive advantage justifies the implementation of the rules set out in the DMA.
It is not just the DMA. Lately, the idea of fairness has permeated in a number of debates and regulatory frameworks. There is a shift away from the concept of economic efficiency in thinking towards digital regulation, towards concepts of fairness and equity. While fairness is an explicit objective of the DMA, it is also a cornerstone in the AI Act and EU Data Act, which operate somewhere between ex ante regulatory law and competition law.
With this shift towards the curbing of “unfairness” caused by entrenched market positions and misuse of data requires, as well as elevating fairness as an end-goal in itself, new standards and concepts must be developed as to secure legal certainty. The risks for legal certainty and the coherent application of the law are already being a point of criticism.
It seems that – while often perceived as a universal value – fairness is in actuality not so easily defined, let alone operationalised. This is evident from the debates on the long-standing EU competition law prohibition against “unfair pricing” articulated in Article 102a Treaty on the Functioning of the European Union, which lately have received renewed attention at both Commission and Member states level as result of a string of enforcements against unfair pricing in the pharmaceutical market.
Defining fairness: not an easy task
Although both DMA and TFEU rules on competition entailed in e.g. article 102a TFEU have direct effect throughout the European Union, one cannot ignore the fact that matters such as fairness do rely on cultural-historical elements unique to various member states, although one can strive for a more universal, objective framing of the concept as well, along the lines offered by behavioural and experimental economics, which display a certain “fairness preference” in humans.
Indeed, one could argue that the matter of fairness / unfairness has been one of the main topics investigated in law, philosophy and economics throughout the recorded human history, from writings regarding fair market rules in tablets of Hammurabi, Jewish and Christian teachings, and all the way to modern thinkers such as Immanuel Kant but also Adam Smith who famously noted in Wealth of Nations that “Justice is what holds the edifice together”. This is true however, provided that fairness can be objective, rational, and practicable.
This idea of how to make fairness operational would invite the development of some kind of economic concept of fairness. After all, fairness that is based on socio-cultural values would be subject to change and jurisdiction, while fairness as an economic concept would create some level of stability and predictability. While both forms of fairness (socio-cultural and economic) are intertwined with one another and may be instructive for the law, a “purely” economic concept of fairness is likely perceived as more objective, rational, and practicable. The concept of (un)fairness has proven – nevertheless – a complex and contentious one, in both law and economics. This is evident from the debates surrounding the application of Article 102a Treaty on the Functioning of the European Union (TFEU), which prohibits unfair pricing. Although codified already at the birth of the European Union, there has been many cases, and yet even more legal and economic commentary on “fairness” as a goal for competition law.
Charlotte Carlberg Bärg.
Fairness in digital regulation
The intersection of fairness and competition law has noted a revival with added focus on digital markets and tech giants in the EU and opening of investigations and issuance of hefty fines against tech giants such as e.g., Google and Apple (as well as by a string of other cases against “unfair” pricing in other sectors such as pharma). This fairness thinking has ultimately found its way into ex-ante regulation for the digital sphere.
Regarding the “imbalance between rights and obligations” as the be-all definition of unfairness in DMA, one can quickly observe that almost all business transactions display some element of this “imbalance”, and that much business acumen is built on others doing a less wise choice. Certain situations and markets do necessitate a more careful consideration, mostly in cases where the markets are not contested, due to severe barriers to entry, or other characteristics.
Naturally digital markets do fall in this category, a matter which motivated the creation of DMA in the first place, and designation of certain gatekeepers which occupy a unique position due to their size and access to vast amounts of data, creating insurmountable barriers to entry, and also displaying certain abuse of this market power, in turn enabled by the sheer access and control over data-driven platforms such as Meta, Amazon or Apple.
This does however not mean that its inclusion in ex-ante rules leads to a clear and delineable idea of what is fair and what is unfair. The Digital Markets Act mentions fairness 43 times, and unfairness 18 times. The EU AI Act mentions fair in 17 instances, and the EU Data Act mentions the word fair 28 times. Yet with all this discussion on the concept, these laws and their recitals have only been able to draw the contours of what fairness is: an imbalance in the distribution of allocated value, entrenchment, dependency, and the ability to behave unfairly characterise fairness in the DMA. In the EU AI Act, fairness is interpreted in light of the principles of trustworthy and sound AI. This culminates in seven principles: human agency and oversight; technical robustness and safety; privacy and data governance; transparency; diversity, non-discrimination, and fairness; societal and environmental well-being and accountability. Fairness in the AI Act itself is however merely defined as “[d]iversity, non-discrimination and fairness means that AI systems are developed and used in a way that includes diverse actors and promotes equal access, gender equality and cultural diversity, while avoiding discriminatory impacts and unfair biases that are prohibited by Union or national law.” In the EU Data Act, fairness means equitable sharing of data on a sound basis. The Act also defines some behaviours that are by their nature unfair, similar to hardcore restrictions in EU competition law.
There are still many questions surrounding this topic: are these notions of fairness actually based on a sound economic basis? Economic assessment is notoriously limited in the DMA, yet how fairness is defined in the recitals reflects ideals related to economic efficiency and the merits of competition. Meanwhile, fairness in the AI Act seems to result from a socio-cultural concept, it is about social inclusion and non-discrimination. The EU Data Act again reverts back to fairness like we already know it from EU competition law. The question about whether there is one concept of fairness or multiple remains in the air, as the Court of Judgement of the EU has not yet defined fairness in any of its judgements.
The way forward: can we operationalise fairness?
Fairness can also be able to function as an economic concept, provided fairness can be defined objectively, rationally, and practically. The simplistic claim that “fairness is not an economic concept” is fraught with challenges, as it is a matter of choice and values inherent in rejecting fairness as being an incalculable externality, to be dealt with other disciplines.
Indeed, an immense body of economics have concerned itself with different nuances of fairness and equity. As demonstrated by behavioural and experimental research, both firms and consumers take fairness considerations seriously regarding e.g., pricing. Nevertheless, the concept of fairness and its practical application in law and economics is not without manifest challenges in overcoming the inherent “subjective” element entailed in fairness considerations.
The challenge ahead for both ex ante regulatory approaches but also ex post competition law enforcement lies in how to translate the abstract notion of fairness, and the rather vague framing of fairness in DMA and the AI Act, into practical and workable concepts when dealing with matters such as zero price products, algorithmic collusion and exploitative abuses enabled by network effects and misuse of data.
One possible approach would be to define fairness more alongside traditional consumer welfare standard, and equitable exchange, taking a more holistic view of the interest of the trading parties on the market. Nevertheless, in order to avoid Type I and Type II errors, more hands-on guidelines are needed to make the application of the recently introduced acts a practical and objective endeavour. It is clear that, as the notion of fairness still creates as many questions as answers, further research into the operationalisation of this concept is needed.
Behrang Kianzad and others will further explore questions surrounding fairness in digital markets and beyond in the conference “Making digital markets work for people” on the 7th of February 2025. For anyone interested in the topic, there is still time to register and to submit an abstract for the Call for Papers!
