The first post-Brexit Competition Bill: What to watch for
The UK Parliament is poised to publish its Digital Markets, Competition and Consumer Bill (“the Bill”). The essential point is to place the UK Competition and Markets Authority’s (“CMA”)’s Digital Markets Unit (“DMU”) on a statutory footing, and to set out the evidence standards applicable to it.
But this is not just alphabet soup! As the first major post-Brexit competition law reform, the debate over the Bill and its final form will be a bellwether for the direction of travel of UK competition policy. This post provides core points to watch for in reading and engaging with the new Bill.
1. Designation: Timing and scope. The essential point in the Bill is to alter the law related to designated companies. It will be significant to see whether the evidence standard for designation is based on size alone, or on economic evidence of entrenched market power.
2. Non-digital activities and Adverse Effects on Competition (AECs). One of the most fraught aspects of the bill is to identify a boundary between regulated and non-regulated activities. We are told in the consultation that leveraging can be a legitimate strategy. What about non-digital activity? Sometimes, this competes and needs to be taken into account.
3. AECs and PCIs. It is also not yet clear whether an AEC has to be proven in a relevant product market as opposed to just in some digitally related “activity.” The former is more principled as relevant market doctrines are then applied. This provides a good boundary to intervention in the Competition Act 1998 and it is unclear why the same doctrines would not be apt here. If there is a problem, it is not asking much to state which market it arises in on the basis of evidence. Clarifying that PCIs relate to a relevant market would be very welcome.
4. Is there to be an efficiency defence? In merger review and dominance/monopolization law, there is scope to argue that efficiency is enhanced by large business scale. From this perspective, it is not a problem to have even a monopoly in the short term, provided that rivalry exists over the longer term. This is in fact the spontaneous order of modern tech firms: there are successive generations of winners, although concerns sometimes arise within the period when the company is large, as a monopoly may persist for a period of time. The fundamental question here is whether the evidence should be based on impacts on consumers – in which case, low prices even from large scale would be seen to be positive. By contrast, a vision emphasising business entry would be less concerned about the loss of efficiency from decreased scale, since the point would be to encourage entry and in-market competition, even if this comes at some cost.
5. What will become of the consumer welfare standard? The consultation papers for the proposals outlined a defence based on net consumer benefit. That would reverse the burden of proof. Normally, absent market power, consumer benefit is left to consumers in their purchasing decisions. It will be significant to see whether there is a presumption of harm to consumers, or whether this must be proved as part of the affirmative case against conduct and/or before one of the proposed “Pro-Competitive Interventions” can be made. Requiring no evidence of customer impact is a shaky proposition.
6. How will the Bill prevent rent seeking abuse of the process? If evidence standards are relaxed, capture risks increase. Aspects of current evidence standards try to ensure that capture risks are mitigated by working out end consumer impacts. If business interests alone come to suffice, there is risk in looking only to company welfare. In such a scenario, it would not be necessary to prove a consumer impact (although it would certainly be wise to do so, from both the regulator’s and the complainant’s perspective).
It will be significant to see whether the legislation asks whether consumers are harmed by the business practice. Legislation could potentially also ask for the DMU to ensure that the intervention does not harm them. Reviews and reporting could address this by requiring particular evidence of impacts, and by reviewing them over time in deciding whether to retain a PCI and/or regulation of conduct.
7. Defining interoperability. “Please interoperate” is not a helpful instruction. Instead, it should be accurately specified and put behind a clear foreclosure screen. Historically, indispensability had to be shown in Bronner v Mediaprint under UK law (inherited from the EU) before access duties would arise. In many ways, the essential point of the Bill is to relax this requirement by placing designated companies under duties to justify conduct affecting rivalry, including denials of interoperability and interconnection.
There is a significant question over what evidence gates will be applied to interoperation: a foreclosure standard, as with tying? A denial of contestable share, as with exclusivity? A requirement for new products to be blocked before intervention occurs, as in IMS Health? That would be very interesting, as it is a test that seeks to promote innovation but blocks “me too” products – but at the risk of missing short term monopolistic practices regarding existing products. A bolder position can be found in the USA: sometimes denying access encourages innovation; thus access can harm the consumer (Verizon v Trinko).
8. Is the concern dominance, or monopolization? US law does not use abuse of dominance. Instead, s.2 of the Sherman Act requires (i) market power and (ii) monopolization thereof, interpreted to mean the extension of market power. There is an interesting history to this, as it reflects skepticism as to the merits of regulation of market power as opposed to avoiding it. If the aim of the DMU is to address market power at source, then logically the focus ought to be on changes to market power, rather than the more EU-style regulation of abuse. Seen in this way, it will be important for the law to require evidence of delta on market power as to (i) designation; (ii) conduct remedies and (iii) especially, PCIs. An affirmative requirement to prove market power diminution from regulation, perhaps at a review point, would be very welcome to avoid utility-style perma-regulation.
9. Benchmarking and review. Whatever one’s view on the correct scope of intervention, benchmarking and review will be important to assess the quality of regulation. The magic phrase here is the “Nirvana fallacy” which derives from a famous article by Harold Demsetz, Information and efficiency: Another viewpoint. Demsetz noted three “nirvana” fallacies (this pre-dated the band…):
(i) Greener grass = intervention will fix it with no evidence to that effect,
(ii) People might be different = I see your evidence but… ignore it, and
(iii) Free lunch (better known = costless intervention).
Benchmarking needs to address whether intervention is truly greening the grass, or not, on an evidenced basis – and tracking of intervention costs (including those to the DMU, not least in ongoing monitoring which is very expensive to get right). It will be important to avoid simply stating a problem without also assessing its net-positive regulation (or not) on an evidenced basis.
The evidence here would be of market power and of consumer impacts. This is often burdensome work for regulators and would ideally be given to a neutral third party to stop bias. This would actually help the DMU as external review and monitoring would provide a basis for retiring old and/or non-cost justified interventions.
10. Conduct requirements:
a. Discriminatory Ts and Cs. Definition will be key here. There is a debate in competition law as between (i) discrimination; (ii) tying of products and (iii) exclusivity practices (must buy all from DomCo). Of these, (i) is the vaguest and most dangerous. Genuine competition risks are greater with tying, and greater still with exclusivity (since both practices can deny an entrant contestable share and thus prevent scale up). If scale is not itself a problem, then self-preference/discrimination language should be limited so as to address only those Ts and Cs which truly raise entry barriers. If not aligned to a clear evidence basis, this would allow complainants to go after big platforms simply for their doing well, without regard for impact on competition. If self-preference is kept on a free-standing basis, then it should be behind a gate of proving competitive impact (and ideally consumer impact).
b. Tying and bundling. Both are complex practices depending on context. Approaches to the so-called “one rents” theory will be important: Two tied products A and B — > tying does not itself increase market power because if there is competition in product B, then tying to A with inflated price for A simply decreases sales of B. Sometimes true and sometimes not (depends on whether the two products are bought in direct proportion and with same consumer valuation). However, the core insight is very significant: the whole market, including related markets, should be considered as a whole. Otherwise, some consumer-helpful bundles can be hit (e.g., lower car price but higher service cost = good if not driving much; cheap printer, expensive toner = good for low volume consumer). Same is true of websites. Requiring attention to all related markets would be positive here.
There is a risk of disruptive unbundling claims, similar to disruptive interoperability claims. The classic American case is almost darkly comic: Jefferson Parish v. Hyde, in which a patient argued for a power to unbundle anaesthetic services from operations (denied, unsurprisingly). Surely no one wants the DMU to support the data unbundling equivalent of unbundling an anaesthetic. The trick here is to identify the consumer-facing market (e.g., an operation) and then check if competition is working there. This prevents undue unbundling within the supply chain.
c. Choice remedies in online platforms. Excessive unbundling risk from giving users choice over “their data” but… is it actually their data? If pseudonymised, it may not be. There should be clarity so as to be consistent with new data bill which only regulates where there is a reasonable risk of identity revelation. Any choice remedy should be consistent with this as the two bills are tabled together.
Due process aspects
· There are debates to be had over the review standard involved, e.g., appeal vs review and on what evidence.
· The Competition Bill is aimed at a handful of particular companies. Due process concerns arise from this. At worst, this could become a bill of attainder and that is not desired by anybody.
· A good example of due process issues can be seen across the pond in the fight over whether the US FTC can issue rules. The same points would apply to PCIs and conduct requirements which amount to rulemakeing:
o FTC jurisdiction over unfair methods of competition — > Tenuous argument for rulemaking (obscure and very general reference to power to make “rules” in s.6(g) + old 1973 case of National Petroleum Refiners saying that fuel pump label rules could fall within this).
o If not able to make competition rules (as Supreme Court would likely hold) —> FTC must proceed via determinations with due process protections – specifically, particular evidence of particular harms to competition.
o FTC jurisdiction over unfair practices — > clear authority to pass rules under 1975 reform known as Magnuson-Moss. This allows FTC to provide rules but only subject to specific due process safeguards, notably an obligation to consult and respond to all points raised by affected parties (“hybrid rulemaking”).
· The essential argument for due process here is very strong. As this is not a market investigation, it is not industry wide. It is much closer to Competition Act 1998 investigations, in that the rights and liberties of a particular company are directly implicated (albeit on the basis of a wider marketplace effect). It is one thing to widen the scope of competition law on the basis that proving effects is too hard; it is quite another to remove procedural safeguards around how people are affected by determinations.
· In this regard the PCIs raise concerns. They ought not to be a black list. There are examples of black lists in consumer protection law but they are very limited (e.g., disclaiming negligence liability in relation to personal injury in UCTA/SGSA). A grey list is much more helpful as it allows for evidence to be assessed.
· One practical way to implement this would be to have identified conduct sit behind a presumption, just as is done with consumer protection law. There, a rebuttable presumption of unfairness exists (e.g., ability to vary price after contract forms). It is still open to introduce evidence that the practice is acceptable.
Data
· There is much scope for mischief relating to data under a reform. Actually, this may be one of the areas where PCIs could have a role, e.g., by preventing abuse of GDPR to stop smaller rivals from handling their data (though it should be noted that this is addressed by the existing Competition Act 1998 powers, as in the CMA’s Privacy Sandbox case).
· Considering the role of data in promoting competition, it is notable that the consumer interest is usually in data flowing so as to support ad funded business, and not in control over it. E.g., user 1 went to the BBC website and then looked for a CD. Not exactly private and safeguards can be used to address concerns (e.g., stripping out sensitive data from collection and processing).
· Thus the requirement for consumer choice regarding data is concerning. This does not align with consumer protection law, where the price of a product is not justiciable (ss.61-2 CRA 2015). If data is the price of website access, the same provision ought to apply. There is also a safeguard preventing regulation of free digital services in s.33(5): not regulated unless a minister agrees to regulate a free digital service, doubtless because the risks to the consumer of harm from a free product are low or zero.
· By the same token, competition law ought not to regulate, unless there is clear evidence of a foreclosure risk (which can arise even for a free input).
Departures from and tensions with other areas of English and UK law
· There is a risk of departure from the general position on the consumer. There is no inequality of bargaining power doctrine for B2B contracts (NatWest), no good faith duty (Walford v Miles), no justiciability of price even with the consumer (ss.61-2 CRA), no duty to deal with another business or to give reasons for termination (M&S v Baird), duress narrowly construed (all those Lord Scarman cases inc. Pao On). This is the humble idea that the court does not know better than business, provided that competition is working.
· So what is the logical difference for the DMU? The only logical and consistent answer must be market power. Thus there must always be evidence of it, or the Bill simply overrules a great deal of English contract and UK consumer protection law, and this is not intended.
As the Bill is debated, these and many other areas of contention will doubtless come to the fore.