‘Baysanto’: EU merger clearance of the Bayer/Monsanto confirms continued attention to innovation markets

Last week, the European Commission confirmed that it had cleared the proposed merger between Bayer and Monsanto, the third mega-merger in recent times between seeds / pesticides giants following those between ChemChina and Syngenta, and Dow and DuPont.

In common with the previous mergers in the sector, the Commission has required divestments and commitments by the parties in order to remove problematic areas of overlap.  And these again focussed not only on existing products, but on areas of innovation where both parties had active R&D projects.

The full decision is not yet available, so it is unclear whether the Commission has gone as far in identifying potential ‘innovation spaces’ as in Dow/DuPont, where it used a patent citation analysis to measure ‘innovation output’ (see the relevant section of our Survey of IP/competition law developments from 2017 for more detail).  However, it has identified a number of areas of where innovation needed to be protected before the merger could be approved, including innovation competition in:

  • GM and non-GM traits (i.e. specific plant features traceable to identifiable genes) conferring herbicide tolerance or insect resistance; and
  • herbicides / herbicide systems (i.e. herbicide combined with a trait ensuring herbicide tolerance).
Other areas of innovation (e.g. biological pesticides) were not considered to raise competition concerns.

Relevant commitments offered by Bayer include divestment of its vegetable seed and broadacre (i.e. large-scale agricultural) seeds and traits businesses, including the associated R&D units in each case, as well as three specific lines of research for non-selective herbicides (to be divested to BASF, subject to a separate merger approval).  Bayer also committed to license its digital agriculture offering and pipeline to BASF, allowing it to replicate Bayer’s position in the EEA (where Monsanto was otherwise the main competitor to Bayer’s system).

The increased focus on very early stage product developments is controversial.  In its focus on ‘innovation spaces’, the Dow/DuPont merger traversed new ground in merger analysis, considering product pipelines at a time even before specific products may have been identified.  Our new CLIP of the month, by Andrea Lofaro, Stephen Lewis and Paulo Abecasis is an ideal introduction to the economic questions involved, including whether consolidation is likely to decrease incentives to innovate (as the Commission fears) or whether it may in fact encourage innovation (both by the parties and by others in the market) and enhance the ability to take advantage of innovation that has been carried out.  

Debate will no doubt continue over these questions for some time to come.  In the meantime, Bayer and Monsanto will continue their efforts to obtain clearance for the transaction from other authorities, including the USA.

Race to publish rival SEP & FRAND Codes of Conduct: new CEN-CENELEC working groups established

In October last year we reported on the difficulties that the Commission was facing in drafting its Communication on SEPs, in particular relating to the issues of use-based licensing or chipset licensing (see here).  We also noted that certain industry participants, such as Nokia and Ericsson, had confirmed their intention to establish an industry-wide code on best practices for SEP licensing.

In the subsequent months there have been some significant developments. The Commission’s Communication has been published (here, analysed in an article we wrote for the CIPA Journal here), although there was a notable absence of any specific reference to use-based or chipset licensing. The Nokia-backed proposal for a Code of Conduct has also crystallised into the form of a CEN-CENELEC workshop that kicked off in October 2017 (WS-SEP). 

Perhaps alive to the risk of that workshop producing a Code of Conduct that favoured SEP holders, a rival CEN-CENELEC workshop was set up by ACT (The App Association) and the FSA (Fair Standards Alliance) in February 2018 (WS-SEP2). Although slower off the mark, this second workshop is operating to a faster timetable; both workshops aim to produce a final text of their Code of Conduct by June 2018.

The detailed project plans for each workshop reveal exactly the sorts of differences in approach that one might expect. Both workshops intend their participants to discuss a range of FRAND issues in order to identify best practices from which a Code of Conduct can be developed. However, there are some clear differences in emphasis from each workshop.

WS-SEP focuses on the goal of concluding licence agreements and resolving licensing disputes quickly and efficiently. As part of this, it proposes the establishment of an ‘IoT SEP Licensing Gateway’, describing this as a “process and structure to engage in licensing discussions and resolution in a streamlined or more systematic way”.

On the other hand, WS-SEP2 envisages a broader range of topics being covered, making specific reference to the issues of patent hold-up and licensing without mandatory bundling on a portfolio only basis.

The differences in approach are perhaps most clearly seen in the sections of each workshops’ project plan which deal with what we have previously described as use-based and chipset licensing (here) – topical issues that the Commission did not address directly in its SEP Communication. 

Both workshops recognise the importance of developing guidance on the value of a standardised technology. However, while WS-SEP2 states that “a patent owner cannot seek to exaggerate the value of its patent by focusing on value created by downstream innovators and devices”, WS-SEP warns that “the price of a car does not reflect the value of connectivity in a car, equally the price of a chip may not reflect the value that connectivity, enabled by a chip brings to an end product”.

And though WS-SEP2 intends its guidance to cover “[t]he long history of FRAND licensing at all levels of the supply chain”, by contrast, WS-SEP takes a much narrower approach, proposing discussion on “[t]he appropriate licensing points to efficiently provide access-for-all”. 

The Commission’s decision to avoid taking a stance on either of these issues in its Communication, the value at stake in the light of the IoT and the impact that use-based and chipset licensing could have on royalties paid in new licences meant that these areas were always likely to become a battleground.  

We will be very interested to see the final products of both workshops. However, if the end result is two conflicting Codes of Conduct, they may have little impact on resolving the most contentious FRAND issues. 

Depression delayed: CMA’s paroxetine pay-for-delay case heads to Luxembourg

On 8 March 2018, the Competition Appeal Tribunal (CAT) gave an initial judgment (see here) in the appeals brought by GlaxoSmithKline (GSK) and a number of generic manufacturers against the Competition and Markets Authority’s (CMA) 2016 Paroxetine decision (see here and here).   As explained further below, the CAT has (in a move which perhaps goes against the prevailing zeitgeist) not reached final conclusions on the appeals, but has rather referred a number of questions to the Court of Justice in Luxembourg.

Background

The infringements identified by the CMA in its Paroxetine decision arose out of three patent settlement agreements made in 2001 and 2002 between GSK and various generic manufacturers of paroxetine.  

Paroxetine is an anti-depressant (selective serotonin reuptake inhibitor “SSRI”), marketed by GSK under the brand name “Seroxat”, which during the infringement period was one of its highest selling products, accounting for £71.6 million (10% of revenue) in 2001.

The CMA’s investigation into paroxetine was the first UK case to grapple with the contentious area of patent settlement agreements which limit generic companies’ ability to bring their own product to market.  The investigation was formally launched by the OFT in 2011 on the basis of information obtained by the European Commission through its pharmaceutical sector inquiry (2009). 

Following a significant further period of investigation, the CMA in 2016 issued a decision fining GSK, Alpharma and Generics (UK) a total of £44.99 million for agreeing to delay the entry of generic paroxetine in breach of Chapter I and/or Article 101 TFEU.  GSK received the largest penalty, being fined £37,606,275 for its parallel infringement of the Chapter II prohibition:

  • The CMA found that between 2001 and 2004 GSK agreed to make payments and other value transfers of over £50 million to generic suppliers of paroxetine; this amounted to a restriction of competition by object and/or effect. 

  • The CMA also found that GSK’s conduct induced generic providers to delay their efforts to independently enter the UK paroxetine market, abusing its dominant position in breach of Chapter II of the Competition Act.

The CAT Judgment 

The essential element of the CAT’s March 2018 judgment is the reference of five issues to the CJEU for a preliminary ruling.  As things stand, the exact text of the questions has not been formulated (and will no doubt be the subject of fierce debate among the parties).  However, the issues referred will relate to the following areas: 

  1. Potential competition: whether the existence of an interim injunction against generic manufacturers was an insurmountable barrier for entering the market.

    The provisional view of the Tribunal was to find that it was not (since, for example, it could have been discharged). However, it decided to refer a question to the CJEU as the question was similar to one raised in Lundbeck’s appeal (see here). This reference raises issues not dissimilar to those at play in the recent Roche judgment of the CJEU, which considered whether potentially unlawful products could be viewed as potential competitors (see here).

  2. Restriction of competition by object: when the strength of a patent is uncertain, does a transfer of value from the originator to a generic of an amount substantially greater than avoided litigation costs, under a settlement agreement in which the generic company agrees not to enter the market with its generic product and not to challenge the originator’s patent, constitute a restriction by object?

    The CAT emphasised a number of points in connection with this question, including a notable recognition that (a) the uncertainty over patent strength means that a possible outcome of the litigation was that the generic challengers would be held to infringe a valid patent; and (b) an outcome of litigation which upheld the patent should not be viewed as a less competitive outcome than the situation where the patent was overturned.

    A second related question was also identified, which seeks to establish whether a settlement comprising a value transfer which also provides some benefits to consumers (in the form of limited supplies of authorised generic products) should also be viewed as restrictive by object. Again, the preliminary view of the Tribunal was that such limited competitive benefits are not sufficient to draw into question the overall categorisation of the agreement as a by object infringement. If that is correct, it is necessarily a conclusion which will have to be considered in detail for any specific patent settlement agreement, and suggests that a blanket ‘by object’ approach will not be warranted.

  3. Restrictions by effect: in order to show a restriction by effect is it necessary to establish that the counterfactual would have been more competitive? The Appellants argued that the CMA’s Decision did not sufficiently consider the potentially pro-competitive effects of GSK’s agreements with generic manufacturers, for example the savings to the NHS compared to the situation where no authorised generics were on the market.

    The CAT also started to grapple with an issue which has been underplayed in the European decisions to date, namely the relevance of the outcome of the underlying patent litigation. If one realistic outcome of that litigation was success for GSK, the approach proposed by the CMA would be to reduce the test for identifying effects to “the probability of a possibility”. On the other hand, the CAT did not seem to acknowledge the risk of creating a situation where an agreement is considered restrictive of competition by object yet does not possess the requisite degree of probability for an effects finding that is not made out. This is surely an issue that will have to be fully played out before this, and similar cases, are finally resolved. (Damages litigation in relation to patent settlement agreements is likely to bring this issue to the fore, even if the CJEU elects to side-step the question.)

  4. The correct approach to defining the relevant product market: is the relevant product market paroxetine or all SSRIs? While the Tribunal supported the CMA’s finding of dominance on the basis of a market limited to paroxetine, it criticised its reasoning. It considered the CMA to have taken an overly narrow approach to market definition, based on the impact of generic entry on the price of paroxetine - following the Commission’s ‘natural events’ approach used in its AstraZeneca Decision (2005) (see here). However, the CAT supported the view that from the time when there were potential generic entrants, the market was limited to paroxetine and its generics. It recognised that this preliminary view, which suggests a significant change in the relevant market over time (despite no suggestion that the view of prescribing professionals was subject to a similar change) was a departure from existing case law. In the authors’ view, this approach will lead to legal uncertainty and, more importantly, inappropriately substitutes an analysis based on perceived competitive constraints for an assessment based principally on objective demand factors. As the CAT itself notes, this approach would suggest successful drugs will almost always be found to constitute a distinct market at least from the time when generic entry becomes likely. As the Tribunal notes, this approach would amount to a material change to the “IP bargain” which “might adversely affect the economic purpose of patent legislation”.

    Nevertheless, in making the reference on this point, the CAT has flagged some significant issues which should weigh in the CJEU’s eventual analysis.

  5. Abuse: are potential benefits to the NHS relevant to the assessment of whether GSK had abused a dominant position by entering into the agreements?

    The final question on abuse is limited to an allusion back to the questions referred in relation to the object and effect of anti-competitive agreements. The central issue is again the question of whether the limited pro-competitive benefits derived from the presence of the generic companies as distributors of an authorised generic product are sufficient to undermine the main finding as anti-competitive effects.

Although the judgment is provisional in nature, there is much to absorb. We will report further when the agreed text of the questions to be referred has been published.

The BKA’s Facebook investigation: new frontier or regulatory overreach?

At the end of 2017, the German competition authority (BKA) provisionally concluded that certain Facebook polices in relation to users’ data are an abuse of its dominant position in the German market for social networks.  It published an accompanying position paper

The BKA’s provisional dominance finding takes into account Facebook’s 30 million monthly users in Germany (with 23 million using Facebook on a daily basis), as well as the allegedly high barriers to entry for new social networking platforms (a point to which we return below).  

It considers the terms and conditions Facebook imposes on its users to be abusive, requiring them to choose between accepting “the whole Facebook package”, including an extensive disclosure of personal data, or not using Facebook at all. 
 
Specifically, the BKA takes issue with the requirement for users to accept Facebook’s right to collect data from third party websites: data which is made available to Facebook by operators that have embedded the Facebook ‘like’ or ‘login’ options.  Data is collected by Facebook even where users do not click on such options, and is then associated with the user’s Facebook account.  According to Cliqz, Facebook’s reach may stretch to over 25% of all websites. 

The BKA considers that users could not meaningfully consent to Facebook’s data processing requirements, as they have no alternative but to consent if they wish to access Facebook’s platform.  The BKA characterises this as "exploitative business terms”. 

As we have previously noted, the investigation is a test case on the interplay between big data, consumer protection and competition law (here and here). But is this investigation an opening salvo in a new frontier for competition enforcement, or more of a modish dalliance with the hipster zeitgeist?

It certainly concerns an unusual alleged form of abuse, since it does not appear to depend in any meaningful sense on Facebook’s dominance (save perhaps in the extent of the data collection).  Nor does it readily seem to fulfil the classical requirement of an abuse which uses “methods different from those governing normal competition [which …] has the effect of hindering the maintenance of the degree of competition existing in the market or the growth of that competition” (Case 85/76, Hoffmann La Roche).

It is not uncommon for investigations which challenge the boundaries of antitrust to be relatively clear about the potential harm to competitors, but much less clear about the harm to consumers or the competitive process.  This case is rather the converse – with its most controversial aspect perhaps being the BKA’s theory of consumer harm, which is based on users’ loss of control of their data: “Facebook offers its service for free.  Its users therefore do not suffer a direct financial loss from the fact that Facebook uses exploitative business terms.  The damage for the users lies in a loss of control: they are no longer able to control how their personal data are used”.  This theory necessarily assumes that users are sufficiently tied into the platform to be unwilling or unable to ‘click away’ and select an alternative messaging service.  Indeed, in Facebook’s public statement about the recent development in the BKA’s investigation, it strongly emphasised its lack of market power, which is both a pre-condition for any abuse finding and in this case is closely tied to the nature of the abuse itself.

The BKA’s position paper seeks to redress this apparent focus on consumers, noting the high economic value of the data gathered by Facebook, and its relevance for targeted advertising which in turn makes Facebook more attractive to advertisers, a concept referred to as “identity-based network effects”.  The interest in online advertising is of course not limited to the BKA, but is an increasingly hot topic in competition policy generally – see for example the ongoing French competition authority sector inquiry (see here and here). 

It is also questionable how the case can be reconciled with the Court of Justice’s more orthodox position in Case 238/05 Asnef-Equifax (here): “any possible issues relating to the sensitivity of personal data are not, as such, a matter for competition law”.  The BKA’s answer will have to be that its investigation does not relate to the sensitivity of personal data “as such”, but rather to Facebook’s use of the data.

In terms of next steps, Facebook can defend its positon and/or offer solutions.  A final decision in expected in the summer.