A model for a participative approach to digital competition regulation

Executive summary

Digital competition regulations often put in place participative approaches to ensure competition in digital markets. The participative approach aims to involve regulated firms, stakeholders and regulators in the design of compliance measures. The approach is particularly relevant in complex and fast-evolving digital markets, where whole industries often depend on the behaviours of the regulated firms. The participative approach enables stakeholders and regulated firms to design compliance measures that are optimal for all because they ensure legal certainty for regulated firms, save time for regulators and take into account the views of stakeholders.

However, the participative approach is subject to regulatory capture. The regulated firms and stakeholders might try to promote their interests to the regulator. This could result in endless discussions at best, and the adoption of inappropriate solutions following intense lobbying at worst.

A governance model is necessary to ensure that the participative approach works without risks of regulatory capture. The model should define clearly each participant's role, duties and rights. There should be: 1) equal and transparent access of all stakeholders to the dialogue; 2) the presentation of tangible and evidence-based solutions from stakeholders and regulated firms; 3) public decisions from the regulator that contain assessments of the proposed solutions, with guidance to clarify rules; and 4) compliance measures proposed by the regulated firm in line with the guidance. The model should provide an assessment framework for the proposed solutions to identify the most effective. The assessment should rely on the principle of proportionality to assess whether the proposed compliance measure is proportionate, to ensure the effectiveness of the regulation. Finally, the model should safeguard against regulatory capture thanks to transparency rules and external monitoring.

1 Introduction

The last three years have seen the European Union, Germany, the United Kingdom and the United States, among others, adopt or draft laws to regulate competition in digital markets that are dominated by a few large digital platforms (OECD, 2022). Most have opted for a participative approach, with regulators involving stakeholders – for example trade associations and civil society groups – in the design of compliance measures[1]. This approach is relevant in digital markets because they are complex and evolving, and whole industries often depend on the business practices of regulated firms.

Nevertheless, regulated firms can still lack an understanding of which practices they can implement under competition laws or new regulations, because of the novelty of those rules or room for interpretation. Also, competition authorities often let regulated firms propose compliance solutions in response to an order or rule. They thus do not specify how the regulated firms should comply. Proposed solutions impact stakeholders directly, but competition authorities only intervene after implementation to assess whether the solution complies or not with the order or rule. Intervention can then take time, out of step with the rapid development of digital markets, sometimes causing irreparable damage to the economy as the practice can give the regulated firm an insurmountable competitive edge over affected stakeholders, without being effective in terms of compliance with the order or rule. In the face of these uncertainties, the participative approach helps stakeholders and regulated firms to design compliance measures that are acceptable for all, offering legal certainty to regulated firms, saving time for the regulators and taking into account the views of stakeholders.

However, a participative approach has major drawbacks. Stakeholders and regulated firms might have different views. Stakeholders might also try to influence regulators: so-called “regulatory capture” (Stigler, 1971). In that situation, the approach might, at best, lead to endless discussions and impossible solutions and, at worst, to the adoption of inappropriate solutions following intense lobbying by interested participants.

While recently adopted and proposed laws envisage a participative approach, they do not propose a governance model setting out how that approach will work. This policy contribution proposes a governance model that would define the roles of the parties and regulators based on clear participative rules. It would also provide a framework for assessment of the proposed solutions, to identify the most effective. Last, it would safeguard against regulatory capture thanks to transparency and external monitoring.

2 Participative approach templates

The EU and the UK have opted for a participative approach in regulation and some antitrust cases.

2.1 The EU approach

The EU adopted in July 2022 the Digital Markets Act (DMA)[2]. This imposes a list of ex-ante positive and negative obligations on some large online platforms considered ‘gatekeepers’ in some core platform services because they control access to end-users and business users (Recital 6 and Articles 2 and 3)[3]. Obligations under the DMA are divided between those subject to specifications, such as the ban on the promotion of a gatekeepers’ own services over those of rivals in so-called self-preferencing[4], or interoperability requirements between messaging services (DMA Articles 6 and 7) and those not subject to specifications, such as the prohibition on combining personal data without user consent (Article 5). Gatekeepers must self-comply. In other words, they must establish and implement compliance measures.

However, the regulation allows for a regulatory dialogue between the gatekeepers and the European Commission, which is the sole enforcer. The aim of the dialogue is to specify some compliance measures for obligations that are subject to specifications or, in case of circumvention, with all obligations (Recital 65).

In practice, the dialogue is at the core of the Commission’s approach to ensure the effectiveness of the regulation. The Commission is already informally and formally engaged in discussions with all relevant stakeholders on specific provisions[5], such as the DMA ban on self-preferencing (Article 6(5)). Discussions have shown that, firstly, potential gatekeepers and stakeholders have different views on compliance solutions. Secondly, stakeholders, some of whom are plaintiffs in competition cases against potential gatekeepers, have suggested compliance measures that promote their own interests, or require potential gatekeepers to remove some of their services. For example, some shopping comparator rivals proposed that Google remove its comparison shopping service from its search engine, a request that promotes their interests, but that Google is unlikely to follow. However, it remains unclear how the Commission will specify measures, and how potential gatekeepers should implement them.

2.2 The UK approach

The UK government proposed a competition regime for digital markets in July 2021, following recommendations from competition experts (Digital Competition Expert Panel, 2019) and the UK Competition and Markets Authority (CMA)[6]. The regime imposes codes of conduct on large online platforms considered to have “strategic market status” (SMS) because they have a significant intermediary power in some digital markets. The codes are specific to each SMS and contain provisions similar to the DMA. The Digital Markets Unit (DMU) within the CMA is in charge of enforcing the regulation and ensuring “that firms with SMS comply with the regime by combining a participative approach with the use of formal powers” (HM Government, 2021). The participative approach consists of dialogue with stakeholders through advice and informal engagement. The DMU can also open formal investigations to ensure compliance with the codes. However, at time of writing, neither the UK government nor the DMU have outlined a participative approach model.

2.3 Competition cases

The EU and the UK have taken a participative approach in antitrust cases in digital markets to secure commitments.

The UK CMA opted for a participative approach to resolve competition issues in the Google Privacy Sandbox case, in which Google committed to remove third-party cookies that track web-user activity in its Google Chrome browser. The removal would prevent advertisers and publishers from tracking users’ activity across the web in order to offer personalised advertising. Such tracking also gave a competitive advantage to Google, which already has a strong position in the online advertising industry[7]. The CMA opened the investigation in 2021 and closed it a year later, having secured commitments from Google that address competition and privacy issues. As part of the process, the CMA consulted with stakeholders on the proposed commitments with a view to improving them. The CMA then worked with Google to modify the commitments in order to address issues raised by stakeholders, before Google removes third-party cookies from its browser.

The European Commission followed a similar approach in the Amazon Marketplace and Amazon Buy Box cases in December 2022, concerning the alleged promotion of Amazon’s logistics services and the use by Amazon of non-public data of business users to create Amazon products[8]. The Commission also consulted with stakeholders on the commitments offered by Amazon, and then worked with Amazon to resolve competition concerns raised by stakeholders.

However, consultations  only enable stakeholders to provide feedback on proposed commitments after the firm has made them and only in view of assessing them, rather than to propose alternative solutions. The participative approach thus consists of consultations rather than a dialogue.

3 The proposed governance model

The model should contain three main elements: participative rules, an assessment framework and safeguards against regulatory capture.

3.1 Participative rules

The participative approach involves three main participants: the regulated firm, the stakeholders and the regulator. Under an asymmetric regulation like the DMA, the regulated firm is the ‘loser’ that must demonstrate compliance to the regulator. Stakeholders, which are often competitors, are the beneficiaries of the rules.

A dialogue brings together different parties with different powers and roles: the regulated firm has the power to comply with the rules. In contrast, the stakeholders are able to influence how the regulated firm should comply. Finally, the regulator is the arbiter that mediates the dialogue, and monitors and enforces compliance.

The dialogue enables the bargaining of power between the regulated firm and stakeholders to be recalibrated. Stakeholders are often in an inferior bargaining position because of the significant power of the regulated firm. However, some stakeholders might seek to leverage the dialogue to promote their own interests. In contrast, the regulated firm is likely to defend its interests to minimise the costs implied by the regulation. Consequently, the regulated firm and stakeholders might propose different and diverging solutions. While the risk of divergence is real and might reduce the benefits of dialogue, dialogue is a better approach than a setting in which interactions occur only between the regulator and regulated firms, which could mean that the regulator collects views only from regulated firms, which have strong incentives to promote or protect their interests, leading to a risk of biased conclusions from the regulator.

By contrast, dialogue between three sets of participant means that the regulator collects a range of views, minimising the risk of skewed conclusions. While views from stakeholders might not necessarily make information perfect, as stakeholders do not have access to information hidden by the regulated firms, their views might highlight problems and solutions that regulated firms are unlikely to share with the regulator. Moreover, dialogue increases trust in, and credibility and legitimacy of, the regulator as acting in the public interest, not in the interests of regulated firms or stakeholders.

In the EU, a participative approach through stakeholder consultation for the preparation of policies was proposed already in the 2021 better regulation guidelines (European Commission, 2021). The four general principles for consultations are: (1) participation through an inclusive approach involving a wide variety of stakeholders; (2) openness and accountability by making the consultation process transparent; (3) effectiveness by consulting stakeholders on time; and (4) coherence by ensuring the consistency of consultation process.

These general principles should be the basis of all stakeholder consultations. However, specific participative rules are required to ensure that the dialogue leads to the development of appropriate compliance solutions.

The participative rules should thus contain the following provisions to define the role of each participant, their duties and rights:

1. The regulated firm and stakeholders must have the same opportunities and speaking time to propose solutions in the context of formal and informal discussions, to ensure equal and transparent access to the dialogue. Formal discussions should be public, and informal closed-door discussions (eg meetings, calls) should lead to publication on the relevant website of at least the name of the participants and their affiliation, summary minutes, non-confidential documents and time of the discussion.
2. The regulated firm and stakeholders must engage constructively by presenting tangible and evidence-based solutions to avoid undue rent-seeking. The proposed non-confidential documents and solutions should be publicly available on the relevant website to ensure external monitoring.
3. The regulator must consider each proposition equally and issue a public decision on the proposed solutions, to ensure impartiality and accountability. The decision should contain a reasoned assessment of the proposed solutions and guidance based on the assessment, to clarify rules when appropriate.
4. The regulated firm must propose compliance measures in line with the guidance. The regulated firm and the regulator might seek to make these compliance measures legally binding through a commitment to ensure legal certainty to the regulated firm and save time for the regulator. In that case, the regulator should invite stakeholders to comment on the commitment. Finally, following the consultations, the regulator should publish a decision accepting the compliance measures, or should work with the regulated firm to improve commitments to address concerns raised by stakeholders.

These participative rules would ensure cost-effectiveness by providing an incentive for gatekeepers and stakeholders to cooperate in order to avoid costly and time-consuming legal challenges before the regulators and courts. In particular, regulated firms that participate constructively benefit by escaping fines and reputational damage that would occur in cases of non-compliance. Stakeholders benefit from timely resolution, quick implementation and a cordial relationship with gatekeepers that avoids friction and mistrust. Finally, the regulators save administrative costs, human resources and time. A participative approach through dialogue is more cost-effective than an approach without dialogue, such as in antitrust investigations without commitments. In the latter, regulators and stakeholders are confronted with lengthy and complex legal challenges, with the burden of showing non-compliance on the part of the regulator, and with the risk of losing before the court after several years of proceedings that will not guarantee timely and effective compliance.

Nevertheless, it would be naïve to believe that dialogue will always work. Divergences are inevitable, the cost of compliance for the regulated firm might exceed the benefits of non-compliance, or the regulation might require clarification from the court. In these situations, non-cooperation and non-compliance are likely. Accordingly, the regulator would have to use formal powers to constrain the regulated firm to comply. For example, the DMA imposes fines, potential divestiture and the prohibition of mergers in case of non-compliance following a market investigation (Articles 18 and 30).

3.2 Assessment framework

The DMA mentions explicitly that compliance measures should be ‘effective’ in achieving the objectives of the regulation, namely contestability – the ability to overcome entry barriers and challenge gatekeepers – and fairness – balanced rights and obligations (Article 8 and Recitals 32 and 33). However, the DMA does not define what ‘effectiveness’ means.

Two interpretations are possible. First, effectiveness means gatekeepers must implement compliance measures in line with the rules. In that case, implementation self-certifies compliance. Second, effectiveness means gatekeepers must comply in a way that positively affects contestability and fairness. In that case, the implementation must have an effect.

Some academics argue for the second option, recommending that gatekeepers should prove the effects of the change in their annual compliance reports, for example by showing that users have switched to an alternative provider (de Streel et al, 2022).

It is doubtful whether this is the correct interpretation of ‘effectiveness’ for three main reasons. First, the DMA requires gatekeepers to self-comply with the regulation by design, but not that their compliance positively affects contestability and fairness (Recital 65). Second, the demonstration of the effects is not evidence of compliance or non-compliance. Indeed, the absence of effect, such as the absence of user switching, might not mean non-compliance, but just that users prefer to use the service provided by gatekeepers. Last, the DMA is not about an effect-based analysis, but about compliance with per-se positive and negative obligations. Gatekeepers cannot prove that their business practices are welfare-enhancing by showing pro-competitive effects (Recitals 11 and 67, Article 10)[9]. It would then be contradictory that the DMA requires gatekeepers to show the effects on contestability and fairness, but not the pro-competitive effects of their business practices.

Therefore, it follows that the correct interpretation of ‘effectiveness’ means the demonstration of the implementation of compliance measures in line with the rules. However, this interpretation enables gatekeepers to implement compliance measures that do not impact the market.

For instance, the Dutch Authority for Consumers and Markets (ACM), in a December 2021 decision, requires Apple to allow alternative payment systems for dating apps in its Apple app store, to bypass the Apple 30 percent commission fee on in-app purchases[10]. In response, Apple now allows alternative payment systems, but charges them a 27 percent commission fee for paid apps and in-app purchases[11]. Dating apps must also add the cost of payment processing. It follows that no dating apps in the Netherlands use alternative payment systems because they get no benefits from doing so, as the total cost exceeds Apple’s 30 percent commission fee. However, Apple is fully compliant, as the ACM order only requires it to allow alternative payments, not to drop its commission fees. The DMA also requires gatekeepers to allow alternative payment systems, and does not prevent Apple from imposing a commission fee (Recital 43 and Article 5(7)). Accordingly, Apple’s compliance measure in the Netherlands is likely to also be DMA-compliant, but will likely have no beneficial impact for developers. Even if the DMA were to request gatekeepers to demonstrate the effects on contestability and fairness, Apple could argue that there is no effect because developers do not want to use alternative payment systems. The reason would be not because of non-compliance but because they do not want to incur the total costs of the commission fee and payment processing.

In practice, several compliance measures will likely be effective, in line with the rules. The question then is which solution is most effective? Granted that an effect-based assessment is unlikely, the Commission should use another metric to identify it. The DMA allows assessment of compliance measures under the principle of proportionality. This implies that no alternative and less-restrictive measures would achieve the same outcome while safeguarding public order, privacy and security, and preventing malicious practices (Recitals 29 and 35). It follows that the most effective solution is that which achieves the desired policy outcome with the least-restrictive measures for gatekeepers and stakeholders.

Assessment of compliance measures could rely on three steps:

1. The assessment should first identify clearly the desired policy outcome. For instance, is the objective of requiring access to alternative payments to open app stores to third parties, or to circumvent the gatekeeper’s commission fee? The assessment should thus provide an interpretation of the provision.
2. The assessment should then consider if the restrictive measures are necessary. For instance, is a commission fee levied on alternative payment systems necessary to preserve investment in app stores? The assessment should thus carefully consider them, compared to alternative options that achieve the same policy outcome while safeguarding the interest at stake. The comparison should rely on quantitative or qualitative metrics, such as the pricing options. If there is a less-restrictive alternative option that preserves the interest, it means that the restrictive measure is not necessary.
3. The assessment should finally consider the proportionality of the restrictive measures against quantifiable and qualifiable metrics, such as compliance costs. For instance, is a 27 percent commission fee levied on alternative payment systems proportionate for achieving the desired policy outcome considering the compliance costs for stakeholders? The assessment should use a proportionality analysis based on a difference-in-difference approach. It should consider the costs of compliance with the policy compared to those in the initial scenario without the policy. If the costs of compliance with the new policy exceed compliance costs in the initial scenario without it, it will be unprofitable for stakeholders to apply the new policy. The assessment should conclude that the new policy is not proportionate because no rational economic actor will use it. However, the assessment should only reach a conclusion on proportionality. It should remain at the discretion of the Regulated firm to define the means to achieve proportionality, as the regulator does not have the remit to define price or any other norms.

In this assessment, the absence of effect-based analysis does not mean that economists do not play a role. On the contrary, they play an important role in defining the methodologies and qualitative and quantifiable metrics used to assess the necessity and proportionality of restrictive measures, such as by relying on various cost-benefit analyses and difference-in-difference approaches.

Moreover, the DMA explicitly prohibits gatekeepers from engaging in any behaviour that undermines effective compliance with the obligations through contractual, commercial or technical means, and behavioural techniques or interface design (Article 13). The assessment should consider in particular behavioural techniques, as restrictions might have behavioural components. For instance, a warning message when users use an alternative payment system might dissuade them from using it. The assessment should rely on behavioural economic studies. For instance, several academic studies on the EU general data protection regulation (GDPR) consent requirement found that several providers use deceptive or malicious practices to steer users into consenting to data collection (Nouwens et al, 2020). The assessment of provisions that require consent (DMA Articles 5(2), 5(9) and 6(10)) should thus consider these behavioural studies carefully to ensure that users consent freely. Box 1 illustrates the assessment using the requirement to open app stores to alternative payment systems.

3.3 Safeguards against regulatory capture

The main problem of a participative model is the risk of regulatory capture by interest groups. The regulated firms and stakeholders will try to defend and promote their own interests by presenting only evidence that is favourable to them. They might try to influence the desired policy outcome and present compliance measures that are not necessary to achieve it. Moreover, European and national policymakers might also pressure the regulator to use the DMA to achieve industrial-policy objectives, such as the promotion of national champions by imposing undue restrictions on gatekeepers. There are two ways to minimise the risks of regulatory capture: transparency and external monitoring.

Transparency requires dialogues and non-confidential materials to be publicly available. The Commission is already transparent. Stakeholder workshops are, for now, publicly available, and all recording and non-confidential materials are available online[12]. However, closed-door dialogues are non-public; materials and identities of participants are not disclosed. This provides room for regulatory capture without any public control. To ensure transparency, the Commission should publish on the relevant website the names of participants, summary minutes and non-confidential documents, as recommended in section 3.1. The literature also proposes that the regulator keeps and publishes statistics on the number of meetings (Shapiro, 2011). Moreover, the webpage should also include the Commission’s public decision that contains the assessment. For instance, the Commission could follow an approach similar to the UK CMA when it undertakes market studies, by presenting the reports with responses received from stakeholders[13]. This will ensure the accessibility of the documents received and the impartiality of the public decision, as the assessment should rely on the documents received. While materials provided by firms and stakeholders should inform the Commission, it should have the discretion to rely on its own information in order to do the assessment.

External monitoring requires that external audits are possible. They are two complementary ways this can be done. First, the Commission could proactively request the assistance of external parties, such as independent experts or national competent authorities, to review and assess stakeholders’ submissions. This will provide an independent view, with the expertise required to assess technical issues. The DMA allows the Commission to request independent external experts, auditors and national competent authorities to assist it in monitoring effective implementation and compliance with the obligations (Article 26). Second, the online availability of non-confidential materials enables external parties to review stakeholder submissions. They should be able to submit comments on the stakeholder submissions to offer independent views to the Commission. They should also be able to comment on the compliance measures proposed by regulated firms.

While transparency and external monitoring can effectively reduce the risks of regulatory capture by ensuring public oversight and the accountability of the regulator, they might not entirely prevent regulatory capture. Other factors can play a significant role in influencing the regulator, such as revolving doors with staff passing from the regulator to industry, including consulting firms, law firms, regulated firms and stakeholders (Shapiro, 2011).

The participative approach through dialogue might even risk exacerbating other factors of regulatory capture, such as information capture, which is when regulated firms seek to overwhelm the regulator by constantly submitting documents or requesting discussions. As a result, the regulator might opt for the dominant view expressed by the submissions due to a cognitive bias that overestimates the significance of the dominant information, or due to the inability of the regulator to consider alternative options as it must give feedback on the submissions received. To address these factors, the literature proposes adequate funding of regulators, and making them attractive with career-development opportunities and compensation that matches the market, with sufficient expert staff from the civil service rather than industry, to minimise revolving doors and information capture. Regulators should also publish metrics showing the achievement of the objectives of the regulation they enforce (Shapiro, 2011).

References

CMA (2020) Online platforms and digital advertising, Market study final report, Competition and Markets Authority, available at https://assets.publishing.service.gov.uk/media/5fa557668fa8f5788db46efc/Final_report_Digital_ALT_TEXT.pdf

Digital Competition Expert Panel (2019) Unlocking digital competition, HM Treasury, available at https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/785547/unlocking_digital_competition_furman_review_web.pdf

de Streel, A., J. Crémer, P. Heidhues, D. Dinielli, G. Kimmelman, G. Monti, R. Podszun, M. Schnitzer and F.M. Scott Morton (2022) ‘Enforcing the Digital Markets Act: Institutional Choices, Compliance, and Antitrust’, Policy Discussion Paper No. 7, Tobin Center for Economic Policy at Yale, available at http://dx.doi.org/10.2139/ssrn.4314848

European Commission (2021) ‘Better Regulation Guidelines’, SWD(2021) 305 final

HM Government (2021) A new pro-competition regime for digital markets, Secretary of State for Digital, Culture, Media & Sport and the Secretary of State for Business, Energy and Industrial Strategy, available at https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1003913/Digital_Competition_Consultation_v2.pdf

HM Treasury (2022) Autumn Statement 2022, available at https://www.gov.uk/government/publications/autumn-statement-2022-documents/autumn-statement-2022-html

Nouwens, M., I. Liccardi, M. Veale, D. Karger and L. Kagal (2020) ‘Dark Patterns after the GDPR: Scraping Consent Pop-ups and Demonstrating their Influence’, Conference on Human Factors in Computing Systems, April 25-30, Honolulu, available at https://dl.acm.org/doi/10.1145/3313831.3376321

OECD (2022a) Analytical note on the G7 inventory of new rules for digital markets, Organisation for Economic Co-operation and Development, available at https://www.oecd.org/competition/analytical-note-on-the-g7-inventory-of-new-rules-for-digital-markets.pdf

Shapiro, S. (2011) ‘The Complexity of Regulatory Capture: Diagnosis, Causality, and Remediation’, Roger Williams University Law Review 102(1)

Stigler, G. (1971) ‘The Theory of Economic Regulation’, The Bell Journal of Economics and Management Science 2(1): 3-21