Competition Law Forum hosts the 5th annual tech antitrust roundtable with Linklaters
On 20th November, the Competition Law Forum and Linklaters hosted a roundtable discussion with the European Commission, CMA, Ofcom, FCA, DRCF and leading economists and lawyers.
The discussion included (i) the evolving landscape of platform regulation in the EU and the role of economics in enforcement of the new regime; and (ii) innovation and contestability in digital markets. The session was held under Chatham House rules so none of the content is attributed to anyone individual. Below is some key messages and a few of the take-aways from our esteemed speakers:
The role of economics in the DMA. The Digital Markets Act (DMA) is a legislative Act aimed at addressing the perceived failure of competition law and underenforcement in dealing with markets characterised by high barriers to entry and significant network effects. The underlying philosophy of DMA is fairness, contestability and open markets. While there is no need for a market definition, designated gatekeepers need to show - in a compliance report - how they meet the obligations in the DMA in light of contestability and fairness. A number of economic tools are relevant to show contestability such as consumer surveys, win/loss analysis, a description of the evolution of barriers to entry and multi-homing to mention a few. There are no specific economic tools to show fairness, but arguable a market that is constable is fair, so showing contestability will lead to a conclusion of fairness.
The relevance of FRAND and standard essential patent in DMA. FRAND is a licensing commitment often used in the context of standard-setting organizations and intellectual property rights. It refers to the terms under which a patent holder agrees to license their technology or patents to others on a fair and reasonable basis, without discriminating against any potential licensee. This framework promotes competition and ensures that essential technologies used in standards are accessible to all interested parties on fair, reasonable and non-discriminatory terms. The question we discussed is whether this form of FRAND commitment is relevant in relation to access to data under Article 6(11) of the DMA given the DMA is more about access to market in non-price terms rather than price and Article 6(12) of DMA, which concerns the terms of access to app stores, search engines and social networks, for similar reasons. Another question which was discussed was whether one could look at a gatekeeper as akin to a standard essential patent holder and look at access price before the gatekeeper became a gatekeeper and after.
Mandating interoperability. The question is whether it is right for a regulator to mandate interoperability. There is no need to mandate interoperability in markets with insignificant network effects and ability to multihome. It becomes more relevant in markets with significant network effects and inability to multihome. Mandating interoperability affects incumbents only and there is a consideration around innovation. There will almost always be a trade-off between mandating interoperability on the one hand and dampening innovation on the other. We discussed whether a distinction between vertical and horizontal interoperability is warranted as these are different forms of interoperability. It is too early to mandate interoperability in AI as we are still at a stage where we are discussing the ethics of AI and early indications are that multiple business models might emerge, and mandating interoperability could threaten this dynamic process given the nascent stage of the market.
AI in the UK and the EU. In the UK, the CMA has proposed a set of competition and consumer principles to guide development of AI foundation models. These include: (i) access to key inputs upstream in model development; (ii) choice and flexibility for businesses deploying models and (iii) empowerment of downstream consumers with transparency about the risks and limitations of content generated, as well as cross-cutting accountability across the value chain. We have to be careful about not dampening the routes to the market, so we should not be too quick to create standards. In the EU, given LLMs/AI are not "core platform services", they will currently be primarily examined through their adjacencies with other services (e.g. search).
Scrutiny and auditing of new technologies: In the UK, there is a debate about auditing algorithms and whether it is really necessary fully to "open the black box", or if algorithms could be effectively audited by examining their inputs and outputs.
Interdisciplinarity. It has become increasingly important to take a interdisciplinary look at markets, so coordination on both a national and international level is essential although challenging at times. Horizon scanning of emerging technologies such as quantum computing is important to be proactive and ready when these technologies become mass-market. While different disciplines can at first blush appear to point in different directions, often when you get into the detail, the conflicts can be resolved by reference to regulators' common goals.
Top tips to takeaway...
• Economics plays a role in the DMA. Although it plays a different role under the DMA than it does under competition law. Under the DMA, economics plays more of a role in the rule making phase i.e. in the specifications of the DMA than in the implementation phase. Economics will also be important for measuring the impact of the DMA.
• FRAND is not easily applied as a principle under the DMA, which is about access to data and other services provided by a designated gatekeeper. It is a commitment which is potentially more about non-discrimination than fairness and easier applied in a price context than a non-pricess context. If a gatekeeper were considered akin to a standard essential patent holder one would look at the access price before the gatekeeper became a gatekeeper - again access to data and other services under the DMA is often on non-price access. But this may well not provide an appropriate framework for assessing FRAND access under the DMA.
• Interoperability is a potentially important remedy in helping to address some issues in digital markets, but interoperability interventions could also do more harm than good if not appropriately designed and targeted, and in particular risk distorting nascent markets where business models are developing in a multiplicity of dynamic ways.
• Regulating AI is a significant global challenge where existential risks rightly need to be prioritised, but competition-based regulation also has an important role to play. The current approach at least in the UK is to "guide" the market to support dynamic competition rather than force particular outcomes, but authorities will not hesitate to intervene - with the right tool - if they consider it necessary.