Divergence ratios after Brexit. Parallel EU/UK merger reviews one year on

In the run-up to Brexit, the CMA had made a prediction that parallel reviews of former EUMR one-stop shop cases could mean “an additional 30 to 50 Phase 1 mergers per year”, potentially leading to “half a dozen or so additional Phase 2 cases”.

In terms of divergent outcomes, back in 2020, Platypus wrote about this risk for deals involving parallel reviews by the CMA and European Commission post-Brexit given the UK regime’s potential to surprise the unwary with its lack of compulsory notification but a muscular policy shift on intervention in foreign-to-foreign deals, alongside its imposition of global hold separates in closed deals. 

The CMA had also weighed in on divergent outcomes, noting that conflict was possible but should be “relatively rare” including because “ … both systems are broadly economics-based, apply similar analytical approaches and have similar legal tests for intervening in a merger” but that “in more extreme cases, however, it could lead to the imposition of inconsistent requirements on businesses, potentially frustrating different jurisdictions' remedies, and even causing the abandonment of the merger … what is important is to minimise the risk of different jurisdictions reaching different outcomes on the basis of the same evidence.”

So what does the parallel merger control landscape look like one year after Brexit? The sample size is small, but are there any emerging trends or lessons we can draw out?

Statistics on parallel review cases 

As set out in the table below, as against the predicted tally of 30-50, parallel EU / UK reviews were opened in 12 cases in 2021, or less than half of the lower band estimate.

That might sound like the Brexit effect was overblown – on the other hand, those 12 represented just shy of 30% of the docket of 46 Phase 1 decisions taken by the CMA within this window. 

As the CMA stretches its legs at the top table of antitrust enforcers on multi-jurisdictional deals, most cases have resulted in a consistent outcome with European Commission decisions, as would be expected.

Of these dozen cases, six were parallel unconditional clearances:

Identical outcome - unconditional Phase 1 clearance

	Case Name	EU and UK status	Sector
1	GRAPHIC PACK. / AR Packaging	Phase 1 clearance	Industrial
2	SK HYNIX / INTEL	Phase 1 clearance	Tech hardware
3	AMD / XILINX	Phase 1 clearance	Tech hardware
4	THERMO FISHER SCIENTIFIC / PPD	Phase 1 clearance	Life sciences
5	ASTRAZENECA / ALEXION PHARMA	Phase 1 clearance	Life sciences
6	IHS MARKIT / CME GROUP (JV)	Phase 1 clearance	Financial services

A further four cases, thus far, have the same intervention outcome:

Identical outcomes other than unconditional Phase 1 clearance

	Case Name	EU and UK Status	Sector
7	S&P GLOBAL / IHS MARKIT	Phase 1 divestment remedies	Financial services
8	IAG / AIR EUROPA	EU Phase 2 abandonment  UK no jurisdiction due to abandonment	Air transport
9	NVIDIA / ARM	Phase 2 abandonment	Tech hardware
10	CARGOTEC / KONESCRANED	Phase 2 ongoing	Industrial

Two of twelve cases resulted in divergent or at least non-identical outcomes.

Non-identical outcomes

	Case Name	EU Status	UK status	Sector
11	FACEBOOK / KUSTOMER	Phase 2 access remedies	Phase 1 clearance	Tech services
12	VEOLIA / SUEZ	Phase 1 remedies	Phase 2 ongoing	Industrial

That’s a divergence ratio of 1 in 6 or 17%. As CMA old-timers will know, that is a non-trivial divergence ratio that is arguably more than “relatively rare”.

But the degree to which Veolia/Suez is divergent – beyond the fact that the deal never went to Phase 2 in Brussels but is in Phase 2 in London – could be debated, not least as the deal involves national or sub-national markets.

Do we have really have the same Kustomer?

Facebook/Kustomer is more interesting. The first and highest profile differential outcome occurred in the acquisition by Meta (formerly Facebook) of Kustomer, a small but growing player in the customer service and support customer relationship management (‘CRM') software market. 

The legal tests for Phase 2 are, as the CMA says, similar, but there is no question within the UK antitrust bar that, as a legal matter, the CMA’s “realistic prospect” is lower than EC’s “serious doubts” threshold for Phase 2 -- especially given the Court of Appeal’s IBA Health ruling that the CMA’s test spans to a threshold of belief as low as “more than fanciful”.  

Nonetheless, the CMA unconditionally cleared the Kustomer deal in Phase 1 in September 2021 knowing full well that the EC had referred the deal to Phase 2 the month before, before the EC ultimately approved the deal subject to remedies.

On its face, as an acquisition by an incumbent of a nascent player in an adjacent market (involving complicated theories of harm), the CMA would have had little to blush about if it had sent the deal for an in-depth review. Indeed, while the EC’s decision is not yet available, it is expected that the CMA and Commission investigated similar theories of harm on markets that were global. The CMA’s decision explains that it considered the risk of: (i) increasing Meta’s data advantage in online advertising; (ii) vertical foreclosure by limiting or degrading access of competing CRM software providers to Meta’s messaging channels; (iii) vertical foreclosure by restricting the ability of other B2C messaging providers to integrate with Kustomer’s CRM software; or (iv) cross-subsidising Kustomer to foreclose competitors. 

So unlike some cases involving different facts in different geographic markets, the EC and CMA ought to have been looking at essentially the same facts on the same market.

Despite the CMA’s recent interventionist reputation in the tech space, it was the EC who adopted a tougher stance, approving the merger in phase 2 only in return for accepting ten-year commitments from Meta that it would continue to offer free and comparable access to Meta’s messaging channels (such as WhatsApp / messenger) for Kustomer’s rivals.  

This result stands in contrast to the EC’s review of Facebook’s acquisition of WhatsApp back in 2014, when that larger acquisition sailed through a Phase 1 review. But regulator attitudes have changed on both sides of the channel since then. Indeed, the transaction’s review history shows the regulatory interest in Big Tech acquisitions: the transaction was referred to the European Commission by the Austrian authority and joined by no less than nine other Member States. At the same time, the Bundeskartellamt in Germany decided not to join Austria’s referral up to the European Commission and has separately initiated proceeding under the German merger control rules. The German Phase 1 inquiry is currently ongoing and will run until 11 February 2022.

The UK result can also be contrasted with the CMA’s Facebook/Giphy inquiry. While each deal would have been assessed on its own merits, it is notable that CMA Phase 1 review of the Kustomer deal, which was notified in July and cleared in September 2021, overlapped entirely with the CMA’s Phase 2 inquiry into Meta’s acquisition of Giphy, which opened in April 2021 and led to the Phase 2 Report unwinding the deal in November 2021. The history of that case dates back to at least the hold separate order of July 2020 and features appeals (on procedural breaches) up to the Court of Appeal and two fining decisions of £50.5m and £1.5m for hold-separate order breaches. A judicial review challenge to the CMA’s substantive decision is before the Competition Appeal Tribunal and due to be heard in late April.

Indeed, the CMA’s rejection of an “open access” remedy proposed by Facebook in its prohibition and unwinding of Facebook/Giphy (subsequently accepted by the Austrian authority in that case), and its public criticism of the EC’s remedies in Google/FitBit, raises a series of interesting Phase 2 hypotheticals. Had there been a Phase 2 adverse finding at UK level to match that of the European Commission, the CMA would on all public evidence in other cases have disfavoured the behavioural remedy the EC has embraced as proportionate to solve non-horizontal concerns. The EC’s approach reflects a “middle-ground” on remedies – intervention but in a way that allows the deal to proceed and preserves its economic logic. Conversely, as the CMA’s new Executive Director for Markets and Mergers recently put it: “the case against behavioural remedies is simple and [empirical]: they are more likely to go wrong,".

As to the actual outcome in the UK, the CMA’s Phase 1 realistic prospect test is of course not, in practice, a duty to refer every deal where the risk of competitive harm just scrapes “non-fanciful” – the CMA has a “wide margin” to exercise judgment. Platypus can only speculate that, within this wide margin, a UK Phase 1 clearance was an attractive policy outcome despite the currents of merger enforcement for GAFA deals for several reasons.

• First, while optically in tension, the CMA’s decision to clear would not of course prejudice the EC’s actual Phase 2 review.
• Second, as to real prejudice, it avoided the risk of direct conflict on Phase 2 judgment calls on substance, including the unattractive scenario of the CMA clearing at Phase 2 while the EC diagnosed Phase 2 SIEC concerns broadly in parallel.
• Third, it avoided the yet more unattractive comity scenario of the CMA finding an SLC and rejecting the very remedies package the EC accepted, or was prepared to accept, and blocking the deal, as per Facebook/Giphy. This would have rendered the entire EC Phase 2 process effectively moot just at a time when all the noises are on agency cooperation.
• Fourth, had the CMA blocked and the EC cleared with access remedies, triggering judicial review, the EC result would have stood in direct comparison to the UK result in terms of proportionality and comity, putting the CMA decision at greater risk.
• Finally, from a resource point of view the outcomes happened to achieve the symmetric result that each side of the Channel had one Meta merger inquiry on the Phase 2 docket.

Given the above policy factors, Platypus awaits the Bundeskartellamt outcome with interest, as its 20 April 2021 joint statement with the CMA and ACCC notes that “dynamic markets … require competition agencies to favour structural over behavioural remedies”, highlighting that the latter create an array of risks (under-enforcement, obsolescence, distortion and monitoring costs).

Remedies 

On Phase 1 remedies, there have been some positive signs for global dealmakers in these early guinea pigs that parallel UK/EC processes can be managed effectively. For instance, S&P and IHS Markit were able to align the remedy package of commodity price reporting agencies to satisfy the EC and CMA within a Phase 1 process (although the CMA remedies process necessarily precedes its final approval decision of the remedy package).  

Conversely, Veolia/Suez, a global deal was subject to review in 18 countries, was able to achieve a Phase 1 remedy outcome in Brussels but not with the CMA. The parties did not offer UK Phase 1 remedies, seeming not to have been willing (or able) to offer remedies to resolve Phase 1 concerns of the CMA across eight overlap sectors. The formal on-the-clock CMA process for offering remedies or ‘undertakings in lieu’ at Phase 1 is arguably more compressed but mitigation strategies for this do exist if there is a strong mutual willingness to achieve such an outcome. However, as at EC level, if the Parties believe that the Phase 1 price is simply too high, they are not obliged to offer any remedies, and can argue their case at Phase 2. 

Process

While not entirely a divergent outcome, in the sense that the deal was abandoned globally, IAG and Air Europa had a bumpy ride as they sought to traverse the end of the EC’s one stop shop. Originally, the EC referred the deal to Phase 2 citing concerns around reduction in competition of air traffic to/from Spain. 

A CMA inquiry was ultimately opened very late into the EC’s Phase 2 investigation, despite this clearly being a Spain-centric deal that the Commission was subjecting to in-depth review. While Platypus would have expected similar analysis to have been conducted by both authorities (with any remedy presumably addressing concerns at both ends of any Spanish/UK route), the CMA decided at a late stage to take investigate matters into its own hands. Though the reasons are not public, one can speculate that this was to aid in enforcement (and potentially design) at the UK end of any remedial package. Ultimately, the deal originally conceived in November 2019 was abandoned by IAG, potentially due to a mix of COVID and competition related concerns.

Untested outcomes

One deal that has been squarely in the cross-hairs of multiple regulators for some time is the just-abandoned NVIDIA/Arm, which saw the UK and EU open in-depth investigations and the US FTC sue to block the transaction. The deal was also subject to review in China, with SAMR officially opening its investigation on 25 January. If that were not enough, the case was the first-ever UK referral to Phase 2 of case on national security grounds in the modern Enterprise Act era (since 2003), in this case, in addition to referral on competition grounds. 

While it was announced yesterday that the deal has been abandoned following concerns from various regulators about its effects on competition in the global semiconductor industry, the question of behavioural remedies to solve non-horizontal concerns in dynamic markets would have been tested (see Meta/Kustomer discussion above). Again, this was a deal that raised vertical issues without any obvious structural cure. 

The dogs that did not bark (in parallel)

Of further interest are candidate cases that were not selected for parallel review by the CMA alongside the EC. Here Platypus refers to its fellow animal blog Rhino and identified that 50% of EC “intervention” cases (i.e. cases where remedies were imposed at Phase 1 or the case was sent to Phase 2) in 2021 were not reviewed by the CMA. This statistic relates to six (out of twelve) cases.

For a number of these cases (Suez/Schwarz, Orange/Telekom Romania, Derichebourg environment/Groupe ecore holding), concerns were only identified in certain Member States and therefore we should not be surprised by the lack of a CMA inquiry: where for instance, the effects of these cases are (almost) exclusively outside the UK. Of course, one of the major benefits of the voluntary UK system is that there is no obligation to file where there is little or no effect on competition in the UK, with parties also able to submit a short and confidential briefing paper to obtain soft comfort that the CMA is unlikely to want to investigate the transaction. 

A similarly innocuous story can explain away Kingspan/Trimo, where the EU Phase 2 review is also ongoing. While this case was notified to EC on 3 January 2021 (therefore after the end of the EU one stop shop on 1 January 2021, it had been referred (under Art. 4(5) EUMR) pre 1.1.21 and so was also grandfathered under the Brexit arrangements -- and the EC is also looking at inter alia UK concerns. 

More interesting is the case of Illumina/Grail, where the CMA has not opened an inquiry to date. This is particularly notable given the theories of harm and opposition from the CMA (and the FTC in the US) caused the abandonment of Illumina’s earlier attempt to acquire PacBio. One hypothesis for the lack of CMA intervention is that the CMA determined that it did not have jurisdiction, despite its notoriously malleable share of supply test, due to the absence of target sales or other creative share of supply hook (see Sabre/Farelogix, Roche/Spark) in the UK. If so, it would be a reminder that the share of supply test does have outer boundaries, elastic though they are.

On that note, it appears likely based on the CMA’s provisional findings in Illumina/PacBio that Illumina had over £100m in UK sales ($650m in EMEA, with a big presence in the UK), which means that this deal would have been caught by the proposed amendment to the share of supply test under the Government’s proposals for reform of competition and consumer policy, which removes the need for an increment to the share of supply, an amendment designed to catch purely ‘vertical’ deals with wide ramifications.

Tentative lessons 

So taking a step back, what tentative lessons can we learn from the first 12 months of parallel UK/EC reviews?

• Substantive outcomes seem in step but unfolding of process hard to predict: for the most part, the authorities’ analyses largely follow a similar path in terms of theories of harm but as the cases show, there is potential for divergent outcomes even where there are global markets and similar theories of harm. In the Kustomer case, divergence was not deal-critical but it would have been bold advice to call the precise Article 22/EC/UK outcomes reached in advance of deal signing. Early engagement with regulators, encouraging alignment between them and careful planning of timelines will be key mitigants to aim for success but the Article 22 position means this is easier said than done, as not only will there be cases where the EC lacks original jurisdiction (requiring Article 22 to result in parallel UK/EC review at all) but under the new guidance, the referring Member State(s) do not themselves need original jurisdiction in order to refer.
• Divergences in substance or timing of remedial outcomes is always a factor: it will always be more challenging to deliver a common Phase 1 clearance with remedies. While we have seen successful remedies packages being negotiated across the UK and EC regulators (S&P/IHS Markit), achieving a commercially acceptable remedy package in both jurisdictions may not be achievable in practice, at least in Phase 1 (Veolia/Suez), when the markets are not the same. The recent imposition of a hold separate order by the CMA in an anticipated deal (National Express/Stagecoach) preventing the divestment of part of the target business designed to pre-empt competition concerns may signal hostility towards at least some variants of “fix-it-first” remedies by the CMA. In addition, the divergent approaches to access remedies for non-horizontal concerns in particular may imply a greater likelihood of a workable remedial outcome at EC level but run a binary “block or clear” risk in parallel in the UK.
• Parallel notifications not always required but skip the CMA at your peril: there are a number of deals subject to intervention at EC level where the parties were able to avoid a formal UK investigation (arguably demonstrating the benefits of the UK’s voluntary approach). Platypus is not privy to the reasons, but the late CMA intervention in the IAG/Air Europa deal cannot have been according to anyone’s master plan. We can expect to see more pre-emptive filings filling up the CMA’s docket, as parties won’t want to run a risk of eleventh hour deal delays.

Caveat emptor

The Financial Times warned back in December that dealmakers should beware as regulators across the Atlantic, in particular the US authorities and the EC, appear to have put aside their differences and entered an era of cooperation. On the other hand, the CMA has since 2020 been viewed in the transatlantic antitrust community as the most worried about regulator globally at present and, thanks in part to the nuances of the UK system, has succeeded in blocking deals (e.g. Sabre/Farelogix, Facebook/Giphy) that its US counterparts could not (or would not). 

The experience of parallel EC/UK reviews thus far include examples of great co-operation and alignment between the two regulators leading to parallel outcomes. However, there is no legal or moral obligation on any regulator to guarantee synchronised process or a parallel outcome and there may be good policy, factual or practical reasons why this is not possible even if the parties play every card correctly.  

Early experience on divergence ratios suggests that parallel outcomes should never be taken for granted and divergent outcomes may not be as “relatively rare” as logic might suggest given the common analytical frameworks (given, in particular, divergent philosophies on remedies in digital deals). And while deal mortality - an abandonment or prohibition -- in both regimes may be perfectly symmetrical and suit the authorities, it is not the parallel ending that the NVIDIA or IAG dealmakers had in mind. 

In Part 2 of this series, Platypus will explore parallel reviews by the UK and US authorities, and consider whether there are any trends to draw out across the Atlantic divide.