The CMA never clears a merger on the basis of behavioural remedies. Such remedies do not deal with the source of the competition problem, they are susceptible to circumvention, they distort markets, and they are difficult to monitor. Everyone knows that. The CMA was so wedded to this principle that it was willing to die in a ditch over it last year, in the tiny cloud gaming market that was part of the Microsoft/Activision deal. Except, it now looks like it will accept two behavioural remedies within a couple of months.
First, two leading housebuilders agreed to let the CMA supervise a small number of house sales on a single site in order to avoid a Phase 2 investigation that would have been wholly disproportionate to the competition concerns. That seems like a sensible outcome (even if the remedy was arguably unnecessary).
More interestingly, it seems that the Vodafone/Three merger is heading towards an unprecedented package of behavioural remedies that would allow the deal to complete and thereby reduce the number of UK mobile network operators from four to three. The final decision is due in December, and the CMA may yet end up blocking the deal, but a bit of Kremlinology on the provisional findings suggests that a behavioural remedy is being very carefully considered.
Such an outcome would seek to preserve the benefits of the merger while mitigating any competitive harm. However, it will not be an easy report to write because there seems to be a mismatch between the problem and the solution. And even if the solution is the right one, it will be difficult to design and implement.
The CMA’s competition findings
The CMA has provisionally found competition concerns that would form the basis of a substantial lessening of competition (SLC) in two markets – (1) the retail mobile market, and (2) the wholesale mobile market in which companies such as Sky, Tesco and Lebara buy capacity on the four current networks to sell to consumers. The CMA’s concerns are significant. For example, the projected lower-bound price rises in the retail market would be 7% for Three customers and 3.8% for Vodafone customers. The projected harm to consumers would be £328 million (and possibly as high as £1.1 billion) per year (before any quality adjustments).
Compare this to the CMA’s Sainsbury’s/Asda decision, which blocked the merger on the basis of GUPPI thresholds of 1.5% (in the fuel market) and 2.5% (in the groceries market). The latter was only that high because the parties had proven large efficiency savings that would be passed on to customers. The CMA did not undertake a full merger simulation in the Sainsbury’s case, but as a rule of thumb, the projected price rise would be somewhere around half of the GUPPI. Therefore, Three’s price rises are perhaps ten times higher than the fuel price rises in that case, although that would ignore any adjustments for increased quality post-merger. Fuel, groceries and mobile network services are all essential items in family budgets, so it is not immediately obvious why they should receive different treatment.
Also, bear in mind that 76% of consumers in the CMA’s survey said they were unwilling to pay more for their mobile network services even in return for higher speeds, and 59% of consumers were unwilling to pay more for more reliability. It may be the case that a regulator should overrule consumers’ focus on prices because it will be better overall for the British economy to have three strong networks, and the CMA itself says that consumer views may change over time, but its findings are that price currently matters hugely to consumers in this market.
Assuming these SLCs are confirmed in the final report, the CMA will have a legal obligation to achieve as comprehensive a solution to the competition concerns and their adverse effects as is reasonable and practicable. The remedy would need to be significant in order to eradicate the large incentive to increase prices. The remedy also needs to match the problem.
Possible remedies
The CMA’s notice of possible remedies rightly doubts the chances of success of the types of structural remedy that has been tried in places like Germany, Ireland, Italy and Austria (i.e. a divestiture of spectrum and assets). The CMA does say it will “explore this option further”, but it lists four reasons why the remedy could never work and zero reasons why it might work. Please see recent discussion of these issues on this blog, which seems pretty consistent with the CMA’s views.
Instead, the CMA proposes to explore the following remedy package:
- Investment commitment: This would take the form of a promise to fulfil the parties’ post-merger business plan within a specified time period (after which time no further monitoring would be needed).
- Retail price control: This would limit the parties’ ability to raise prices and therefore ensure that retail customers are protected during the initial years of network integration and rollout under the investment commitment. The controls might, for example, allow existing customers to ‘roll over’ their existing contract terms for a defined period.
- Wholesale access: This would either ensure that wholesale operators get certain terms or it would preserve a certain amount of capacity for them (or both).
This remedy package would be unprecedented. A promise to invest is not even included in the CMA’s merger remedies guidance in the list of possible remedies. Giving it serious consideration may suggest that the CMA is fundamentally persuaded that the merger is good news for the mobile sector in the UK, despite its rather lukewarm language about its benefits (e.g. “some improvement in various network quality metrics in ways that affect consumer experience, but less than the Parties have claimed”) and its disparagement of the parties’ claims about being sub-scale (e.g. “both of the Parties’ standalone networks are likely to deliver higher network quality than they have claimed”). Admittedly, some of this language may be window-dressing to make sure the CMA gets the strongest possible remedy from of the parties.
Mismatch between the disease and the cure
Before we discuss the complexity of the remedy, there is potentially a principled problem with it. The CMA has (provisionally) found a SLC in the form of a permanent incentive to raise prices (and/or decrease quality). In theory, a price cap can remedy that SLC, but not if it is only in place for a specified period of time. The long-term remedy is therefore the investment commitment. This plan must be sufficiently large and successful to comprehensively (and permanently) offset the incentive to raise prices.
This requires a level of speculation on the CMA’s part that it has never been happy to do in the past. The CMA will need to explain in detail that it is confident that, once the plan has been implemented, retail and wholesale competition in the mobile network market among the three remaining operators will have moved to a new, more competitive level that compensates for expected overall price rises in the market of around 1.5-5% per year. It is not good enough merely to have made the promised investments in network quality; those investments must also be sufficiently rivalry-enhancing to comprehensively offset the competition concerns indefinitely. The permanent incentive to significantly raise prices must have disappeared due to stronger competition.
It is not therefore encouraging that the CMA says it has, “some doubts whether the [joint business plan] would – if delivered – offset the anti-competitive effects identified” (para 14.247). The merging parties will need the CMA’s “doubts” to dissipate over the next couple of months. They have had several years to prepare their best analyses in order to persuade the CMA, so it is not clear what further work can be done. The behavioural remedy is not quite a done deal at this point.
Risky remedy
The CMA’s merger remedies guidance explains the CMA’s approach to an acceptable risk profile for any remedy:
“In evaluating the effectiveness of remedies, the CMA will seek remedies that have a high degree of certainty of achieving their intended effect. Customers or suppliers of merger parties should not bear significant risks that remedies will not have the requisite impact on the SLC or its adverse effects.”
The CMA’s desire for a high degree of certainty is a common reason to discard proposals from the merging parties.
There are some reasons to see the proposed behavioural remedy as highly risky, most of which can be found in the CMA’s own publications, which repeatedly explain why it dislikes complex behavioural remedies. For example:
- Specification risks, e.g. the obligations cannot be specified with sufficient clarity to provide an effective basis for monitoring and compliance. It will be difficult to set the right level for any price cap, and state precisely what conduct will deliver the investment plans. The CMA can insist that £X billion is spent, but it will also need to measure how this money delivers better outcomes for consumers and how it increases rivalry between the three remaining network operators compared to the counterfactual of the parties’ pre-existing investment plans. It seems very likely that today’s plan will need to be altered as time goes on. What happens if costs increase and the parties need to scale back the planned investments? What happens when technology changes? And most importantly, the locations and usage patterns of customers will not stay the same, so what happens to the existing plan as this demand changes? Is the pre-existing business plan going to be sufficient or will it need to be embellished, especially given the CMA’s pessimistic language about it?
- Circumvention risks, e.g. the waterbed effect whereby prices are restricted, so the company reduces the quality of its services instead. The CMA will not want to regulate every aspect of the merged entity’s competitive offering, but it will be difficult to explain how it has minimised the waterbed effect otherwise. If the parties must maintain a certain number of sites, the temptation could be to degrade another aspect of their infrastructure that is not being directly regulated, because the post-merger incentive to degrade is structural and permanent.
- Distortion risks, e.g. the price control and investment commitment might affect all three operators’ incentives to compete for new customers and may warp their pricing decisions. In this case, the remedy would (by definition) push prices artificially low and investment artificially high in a market with only two other competitors. Is it distortionary for Vodafone/Three’s investment programme to be publicly known? How is O2 or EE going to win a customer who is effectively locked into Vodafone/Three with a contract that is artificially cheap? How will new Vodafone/Three customers feel about subsidising those customers (assuming the price control does not apply to new customers)? There are good reasons why regulators try to avoid setting prices. Bear in mind that, when Stansted Airport was released from a price cap that was supposed to protect the airlines, its prices halved.
- Monitoring and enforcement risks, e.g. the regulator will always suffer from a significant information asymmetry. By even raising the possibility of such a complex behavioural remedy, the CMA is paying Ofcom a huge compliment.
There is the significant risk of the investments simply failing to deliver (the CMA itself concludes there are “a number of practical implementation risks”). If the CMA adopts the investment commitments as its long-term remedy in this case, the CMA and Ofcom will own these risks. If the remedy fails to deliver, the regulators will be on the hook as much as the merging parties are.
Remember that the CMA normally rejects behavioural remedies that are significantly simpler and more targeted than this one. For example, a simple licence could have been the behavioural remedy that would have addressed the foreclosure concerns in a small part of the Microsoft/Activision deal.
Duration
It will be difficult to decide the appropriate time period for the price controls. If the CMA chooses a specific time period, e.g. three years, it will need to explain why the permanent incentive to increase prices will cease to exist the day after that period ends. The other option is to ask Ofcom to report on when the price controls can be safely released. This would be a difficult task, and experience shows that the price controls could stay in force for a very long time. The broadcaster, ITV, agreed a temporary price ratchet for advertisers in a 2004 merger, which it seems may never be released. The longer a price control is in place, the more likely that the market congeals around it to the detriment of competition.
The joint business plan extends to 2034. It seems doubtful that the price control can be lifted before the investment commitment is substantially completed and the rivalry-enhancing effects have arrived. The price controls may therefore last until the 2030s. The CMA surely cannot force the parties to roll over customers’ current prices for ten years, so the simple rollover mechanism proposed in the CMA’s remedies notice may at least need to be adjusted for inflation but doing so would then tend to be more likely to distort markets.
It seems that the CMA could be reviewing the remedy in the latter half of the 2030s to decide if it can be released.
Conclusion
The CMA and the merging parties will have a busy couple of months, and there is still a material probability that the deal will simply be blocked at the end of it. However, the CMA seems very open to this complex behavioural remedy package and the parties’ response to the notice of possible remedies suggests that they are too.
The merger may be worth all this effort and risk. The parties may have plausible arguments about driving 5G rollout and even ultimately helping to drive economic growth. Analysts seem to agree that Vodafone and Three are currently sub-scale (although the situation is typically viewed with an investor mindset rather than a consumer mindset). I confess that I simply don’t know whether the merger is, or could be, a good thing for the UK.
This blog post makes the more modest point that the behavioural remedy would tear up the CMA’s rulebook and consign all four network operators, more than one hundred MVNOs, both regulators and 65 million consumers to more than a decade of complex arguments about its specification, implementation and review. Even after that, there will still be debate about whether it was the right decision.