Regulating Africa’s Digital Markets: What to Do, and What Not to Do

CPI Columns Africa

By Folakunmi Pinheiro1

I. Introduction

Earlier this year, the African Union passed the Protocol on Competition Policy to the Agreement establishing the African Continental Free Trade Area (“AfCFTA”) (hereinafter, the “Protocol”).2 The Protocol creates, for the first time, an integrated and unified competition regime for the continent together with an AfCFTA Competition Authority (hereinafter, the “Authority”).3

The Protocol applies to all economic activities by persons or undertakings within or having a significant effect within the AfCFTA and that “conduct with a continental dimension.”4 More specifically, there are provisions on restrictive agreements,5 abuse of dominance,6 mergers and acquisitions,7 and abuse of economic dependence.8 However, the Protocol does not apply where national competition authorities have jurisdiction.9 And it specifically excludes conduct, practices, and agreements related to advancing or fixing the terms and conditions of employment.10

It must be said that the Protocol is a very welcome development. The provisions on extraterritorial application,11 the aforementioned labor-related exclusions, the preservation of the jurisdiction of existing regional competition agencies,12 and the institutional structure of enforcement,13 must be commended. There are, however, various concerns with the Protocol. For instance, it does not include a mechanism to determine which competition authority – national, regional, or continental – would be best-placed to investigate a particular anticompetitive problem. State aid provisions are also surprisingly absent. Regardless, these issues may be resolvable with subsequent regulations and guidelines. A more salient (and perhaps more permanent) issue relates to the abuse of economic dependence provision, which this Article will focus on.

This Article will proceed as follows. The second section will provide a brief overview of the economic dependence provision in Article 11 of the Protocol (hereinafter, the “Provision”) focusing on relevant definitions and the broad prohibition on abuse of economic dependence. In the third section, the Article will discuss the specific list of prohibited practices in the Protocol that apply to gatekeepers and core platforms – very much adopting the same language as the European Union’s Digital Markets Act (DMA).14 Next, in the fourth section, the Article will then comment on some broader reasons why caution must be exercised when borrowing language from the DMA. An alternative approach and proposals on possible next steps will be included in the fifth section by way of concluding remarks.

II. Background

For context, abuse of economic dependence is deemed to exist when suppliers or purchasers are dependent on a particular undertaking or a group of undertakings in such a way that there are no sufficient or reasonable possibilities of switching to third parties.15 This dynamic creates a significant power imbalance between the undertaking or group of undertakings and the countervailing power of the dependent undertakings.16 In order to determine whether economic dependence is present, the market shares and relative strength of the undertaking in question are taken into consideration, as well as the existence (or lack thereof) of alternative solutions and the factors that led to the situation of dependence.17

There are concerns which arise in this regard. First, it is unclear who makes the determination of economic dependence. One may assume the Authority, but this is not explicitly stated. Second, it is unclear how the “relative strength” of the undertaking in question would be ascertained in a way that does not merely duplicate the absence of alternative solutions. Third, it is rather odd that the determination of economic dependence includes a consideration of the factors that led to the situation of dependence; thereby presupposing the existence of dependence.

Regardless, the Provision goes on to explicitly prohibit undertakings or “gatekeepers” from abusing their relative position of economic dependence over a customer or supplier.18 As alluded to above, the Protocol essentially adopts the DMA’s definition for gatekeepers: (i) an undertaking that has a significant impact on the AfCFTA; (ii) operates a “core platform service” (which is not defined in the Protocol) that serves as an important gateway for “business users” (again, undefined) to meet “end users” (also undefined); and (iii) it enjoys an entrenched and durable position in its operations, or it is foreseeable that it will enjoy such a position in the near future.19

The Protocol then provides that the Council of Ministers shall develop regulations to designate undertakings as either gatekeepers or core platforms.20 This is confusing on a number of levels. First, a plain reading suggests that gatekeepers are somehow equivalent to, or interchangeable with, core platforms. Meanwhile, the Protocol already defines gatekeepers (for instance, Apple) as entities that operate a core platform service (for instance, a virtual assistant or a web browser). In other words, the gatekeeper is the undertaking, and the core platform service is the service they operate. In fact, a smaller player can technically operate a core platform service (for instance, Opera, the web browser) and not be a gatekeeper. The point being made here is that an undertaking cannot be a core platform service. It is only when an undertaking is a gatekeeper that it can operate a core platform service. Therefore, it is unclear how an undertaking can be designated as a gatekeeper or a core platform. Second, even if we assume that core platforms could be viewed as a standalone and distinct category of economic entity, like gatekeepers, the general prohibition on abuse of relative position of economic dependence does not extend to core platforms (see the preceding paragraph) and instead only focuses on undertakings or gatekeepers. As such, core platforms would occupy an awkward lacuna where they could, in theory, engage in abuses of economic dependence.

III. Prohibited Practices

In addition to the general prohibition of abuse of economic dependence, the Provision also includes a specific list of conduct or practices, similar to the DMA’s obligations, that (undertakings that have been designated as) gatekeepers or core platforms are prohibited from engaging in.21 Again, we continue to be bedeviled by the disjunctive relationship between the two terms.

Nevertheless, the prohibited practices are (a) imposing price or service parity clauses on business users; (b) imposing anti-steering provisions or otherwise preventing business users from engaging consumers outside of a core platform; (c) using business user data to compete against the business user; (d) self-preferencing of services or products offered by the gatekeeper on a core platform; (e) differentiation in fees or treatment against small and medium enterprises; (f) placing restrictions on the portability of data or other actions that inhibit switching platforms for business and end-users; (g) failing to identify paid ranking as advertising in search results and to allow paid results to exceed organic results on the first results page; (h) combining personal data sourced from different services offered by the gatekeeper; or (i) requiring the pre-installation of gatekeeper applications or services on devices.22

Many issues arise here. First, certain prohibited practices could be justified by efficiencies or other beneficial reasons. In other words, they cannot be based solely on anti-competitive intent, nor do they only lead to unjustifiable anti-competitive effects. For instance, with practice (e), mere differential treatment cannot be considered abusive or anti-competitive and could, in some instances, be justified. Meanwhile, practice (b) ignores the fact that there could be legitimate reasons, like security, to prevent business users from engaging consumers outside of a core platform.

Second, certain practices have tenuous links to tangible competition concerns and consumer welfare. For instance, practice (e) focuses only on different treatment, which is extricable from the competitive forces in a market or whether there has been maximization of consumer welfare. Practice (g) does not state how it is anticompetitive or harmful for a search platform to allow paid results to exceed organic results; this should be permissible as a legitimate business model choice.23 While practice (h) could be anticompetitive, as has been found by the Bundeskartellamt,24 the practice is drafted too broadly in order to include all combinations of data from various sources. However, merely combining data from different sources is not per se harmful. In fact, many consumers may enjoy this function of online platform ecosystems. Similarly, practice (i) appears to overlook the fact that many consumers buy devices because of the specific bouquet of pre-installed applications and services.

Third, specifically regarding practice (c), it is unclear what type of business user data should not be used when competing against a business user. It seems perfectly fine for some business user data to be used when competing against a business user––for instance, publicly available data. Commercially sensitive, private data is obviously more likely to be anticompetitive. The practice is too widely drafted and does not make this distinction. As such, subsequent regulations should narrow the practice to focus exclusively on data that is commercially sensitive and not publicly available.

Fourth, and finally, the prohibited practices might not be applicable to all gatekeepers. For instance, it seems practice (g) relates to search engines, practice (i) to manufacturers of devices and/or operating systems, while practice (c) relates to e-commerce platforms. However, one can envisage a quick fix that could clarify and resolve this problem. The Council of Ministers could release subsequent regulations that state the exact core platform services that the Protocol intends to regulate, similar to Article 2 of the DMA.25 After this, the prohibited practices could be allocated to the gatekeepers that operate those specific core platform services.

It is noteworthy that, like the Protocol, the Provision is a welcome development. It is indeed encouraging that the negotiators seriously considered how competition law and policy should apply in digital markets. More regular conversations regarding competition issues in our digital markets are vital and the salient issues that arise in these conversations should be brought to the attention of African competition authorities. However, as this section hopes to have shown, the manner in which this conversation has been crystallized in the Protocol is of particular concern. Nonetheless, that does not in any way detract from an overall positive sentiment that this conversation is happening.

IV. Broader Concerns

So far, this Article has focused on specific and technical concerns with the Provision. Now, it shall turn to focus on the broader reasons why we should be cautious when transplanting concepts and terms from the DMA into the Protocol, after which this Article will conclude with remarks about an alternative approach and next steps.

First, while the DMA is related to and inspired by competition law concerns, it is intended to be sector-specific regulation that is distinct from competition law.26 It creates a sui generis legal regime for digital markets (more akin to the regulatory regimes for the aviation, electricity, communications, and financial sectors). As such, placing this within a document that is intended to be a sector-generic competition law, and not a sector-specific regulation, might raise interpretation, implementation, and enforcement issues for the Authority. In this regard, it is important to remember that Europe’s competition law provisions exist in the Treaty on the Functioning of the European Union (“TFEU”),27 which is similar to the AfCFTA Agreement and its adjoining protocols on competition, investment, and intellectual property rights. Meanwhile, the DMA is a separate and distinct regulation from the TFEU.

Second, it is also worth pointing out that the obligations in the DMA, some of which have been transplanted into the Protocol’s prohibited practices, are not straightforward to interpret. Their definitions are still very much contentious, which is not a surprise as the definitions were established through lengthy, complex investigations and court proceedings––very often relying on the Court of Justice of the European Union to settle the matter. As such, whether these obligations can constitute clear rules that can be followed in practice remains to be seen.

Third, we must remember that the European Commission has, at least, had the experience of these painstaking investigations and court proceedings to develop some capacity and competence to enforce the provisions of the DMA. And yet, there are those who doubt whether the European Commission has the capacity to enforce the DMA.28 Meanwhile, there are very few instances of an African competition authority conclusively investigating an anticompetitive practice in the digital sector or against online platforms. The closest we have is the Competition Commission of South Africa’s Online Intermediation Platforms Market Inquiry, which raised similar concerns to those found in European investigations and cases.29 There is also the GovChat and WhatsApp investigation in South Africa, which I have written about elsewhere,30 but that investigation is yet to be concluded at the time of writing. Regardless, the overarching point stands: African competition authorities are relatively inexperienced when it comes to enforcing competition law against online platforms. As such, one wonders whether the Authority would have the capacity to implement and enforce these vexed and unsettled provisions in practice.

Fourth, the European Commission benefits from a legal, regulatory, and institutional architecture that is able to support its implementation of the DMA. None of these are provided for in the Protocol. For instance, the DMA is a 66-page document, which includes a 109-paragraph recital and 54 Articles. It governs how gatekeeper status can be reviewed,31 which core platform services are being focused on,32 and how the obligations will be updated,33 or even suspended.34 It also creates a framework for how the European Commission can investigate new services and practices,35 as well as a specific Digital Markets Advisory Committee that the European Commission can consult before decisions are made.36 Interestingly, it even creates specific enforcement tools to enable better monitoring of compliance.37 This will invariably complement the growing body of precedent and decision-making practice that the European Commission can rely on when interpreting the DMA. However, the Protocol does not include this architecture; thereby making the Authority’s task, to enforce these provisions, that much harder.

V. Concluding Remarks

For the foregoing reasons, the inclusion of DMA provisions into the Protocol should have been an exercise managed with caution. That said, it could be argued that it is better to proactively include these provisions into the Protocol now, rather than later when the problems of the digital economy have fully manifested. Proactiveness is generally welcome; although proactiveness does not justify widely drafted provisions or overlooking consumer welfare.

Nevertheless, it would have been more appropriate to include the provisions on the digital economy in a separate and distinct document, like the DMA, rather than including them in the foundational documents of the AfCFTA. These foundational documents would be much harder to amend if it becomes apparent at a later date that the prohibited practices of the Provision are not appropriate for the African context. Furthermore, a more stakeholder-oriented approach, informed by concrete competition law interventions into the African digital economy, would have also been a more appropriate approach to creating the Provisions.

However, we cannot go back in time to follow a proper process. What must be done, as indicated in some instances above, is for the Council of Ministers to add specificity through regulations to avoid confusion and interpretational issues. Alternatively, we could look to the national level for State Parties to create their own duplicate regimes of economic dependence; especially because, as noted above, the Protocol does not apply where national competition authorities have jurisdiction. In creating their own regimes, State Parties could follow the process suggested in the preceding paragraph but must exercise caution against undue fragmentation of the regulatory landscape.

Click here for a PDF version of the article


1 Folakunmi Pinheiro is a Doctoral Candidate at the University of Cambridge. His research focuses on competition law and economic development in African countries. He is also a part-time research assistant at Geradin Partners and has previous experience working with the Nigerian Federal Competition and Consumer Protection Commission.

2 United Nations Economic Commission for Africa, “Deepening the AfCFTA: Celebrating the Adoption of New Protocols on Investment, Intellectual Property Rights and Competition Policy” (UNECA, 2023) https://uneca.org/stories/%28blog%29-deepening-the-afcfta-celebrating-the-adoption-of-new-protocols-on-investment%2C.

3 Article 2(a) of the Protocol.

4 Under Article 1(h) of the Protocol, “Conduct with a Continental Dimension” is defined as “any conduct, practice, merger or agreement that has significant effect on competition in a market of at least two State Parties that do not share the same jurisdiction of the existing regional economic communities.”

5 Articles 6, 7, and 8 of the Protocol.

6 Article 9 of the Protocol.

7 Article 10 of the Protocol.

8 Article 11 of the Protocol.

9 Article 3(2) of the Protocol.

10 Article 4 of the Protocol.

11 Article 3(1) of the Protocol.

12 Article 20(1) of the Protocol. Also, see Preamble, which states, “We, Member States of the African Union … CONSCIOUS of the central role that national and regional competition agencies will continue to play in promoting fair competition and inclusive growth in Intra Africa trade and seeking to support their work through the creation of appropriate institutional mechanisms at the continental level … HAVE AGREED AS FOLLOWS …” (capitalization in original). However, it remains to be seen whether the regional economic communities referred to in Article 20 and the regional competition agencies in the Preamble are the same. One could safely assume that they are referring to the same thing and will mean the same thing in practice. But alternative interpretations are possible. For instance, the economic communities retaining their jurisdiction, as per Article 20, could refer to other non-competition agencies that could have concurrent jurisdiction with the Authority to enforce competition rules. Meanwhile, the regional competition agencies of the Preamble are narrower and explicitly refers to competition enforcers.

13 In particular, the fact that there will be an Executive Director which conducts investigations, whose decisions must be approved by the Board of the Authority, and whose decisions can be subject to appeals to the specialist AfCFTA Competition Tribunal. See, Articles 13-16 of the Protocol.

14 Regulation (EU) 2022/1925 of the European Parliament and of the Council of 14 September 2022

15 Article 11(1) of the Protocol.

16 Ibid.

17 Article 11(2) of the Protocol.

18 Article 11(3) of the Protocol. Interestingly, this is provided that the conduct or abuse in question substantially affects the functioning and structure of competition in the AfCFTA. As such, and as one would expect, not all abuses of economic dependence will meet the threshold for intervention. Moreover, one could argue that the Protocol takes a position––by separating the determination from the prohibition and abuse––that holding a relative position of economic dependence is not per se harmful and illegal. Rather, it is only when such a player abuses that position that they will fall within the purview of the Protocol (very similar to what competition lawyers have become accustomed to with abuses of dominance).

19 Article 1(k) of the Protocol.

20 Article 11(5) of the Protocol.

21 Article 11(4) of the Protocol.

22 Ibid.

23 See, Pablo Ibáñez Colomo, “Product Design and Business Models in EU Antitrust Law” [2021] Competition Policy International (on the interaction between competition law and business models in digital markets)

24 “Bundeskartellamt Prohibits Facebook from Combining User Data from Different Sources” (Federal Cartel Office, 2019) https://bundeskartellamt.de/SharedDocs/Meldung/EN/Pressemitteilungen/2019/07_02_2019_Facebook.html.

25 The DMA specifically states which core platform services it will be focusing on. These are online intermediation services, online search engines, online social networking services, video-sharing platform services, number-independent interpersonal communications services, operating systems, web browsers, virtual assistants, cloud computing services, and online advertising services.

26 Luis MB Cabral and others, “The EU Digital Markets Act: A Report from a Panel of Economic Experts” (Publications Office of the European Union 2021) p. 10; and Zlatina Georgieva, “The Digital Markets Act Proposal of the European Commission: Ex-Ante Regulation, Infused with Competition Principles” (2021) 6 European Papers 25 p. 26.

27 European Union, Consolidated version of the Treaty on the Functioning of the European Union, 13 December 2007, 2008/C 115/01.

28 Christophe Carugati, “With a Little Help from Some Friends: Coordinating Digital Markets Act Enforcement” (Bruegel, 2023) https://www.bruegel.org/blog-post/little-help-some-friends-coordinating-digital-markets-act-enforcement.

29 Competition Commission of South Africa, “Online Intermediation Platforms Market Inquiry” (2022).

30 Folakunmi Pinheiro, “Let’s Talk about GovChat” (Compedia, 2021) https://compedia.substack.com/p/lets-talk-about-govchat.

31 Article 4 of the Digital Markets Act.

32 Article 2(2) of the Digital Markets Act.

33 Article 12 of the Digital Markets Act.

34 Article 9 of the Digital Markets Act.

35 Article 19 of the Digital Markets Act.

36 Article 50 of the Digital Markets Act.

37 See, for instance, Articles 14 and 15 of the Digital Markets Act, whereby gatekeepers must inform the European Commission about intended mergers (called, concentrations in the DMA) and they must also submit independently audited descriptions of techniques they use to profile consumers.