US District Court confirms that Qualcomm must offer RAND SEP licences to rival chipset manufacturers

In California, Judge Koh has granted partial summary judgment in favour of the FTC against Qualcomm, making an order that Qualcomm must license its SEPs to rival chipset manufacturers (such as Intel). 

We explained the background to this judgment and its significance in a previous post. Although the ruling is limited to Qualcomm’s position vis-à-vis competing manufacturers without commenting on the position of other SEP holders, it indicates that FRAND requires ‘licensing to all’ (at least in respect of the rules of ATIS and TIA, which, like ETSI, are members of the Third Generation Partnership Project, the collaboration of standard setting organisations behind the development of standards such as 4G LTE and 3G UMTS). The result could be a shift in the focus point for licensing SEPs in cellular standards from the manufacturers of end devices (handsets) to the manufacturers of chipsets.

The issue

Qualcomm had declared its SEPs as essential to two US standard setting organisations (SSOs), the Alliance for Telecommunications Industry Solutions (ATIS), and the Telecommunications Industry Association (TIA). In return, each SSO required Qualcomm to license its SEPs on RAND terms (note that the exact wording of the ATIS IPR policy and TIA IPR policy differ slightly, although this made no substantive difference in these proceedings).

The FTC alleged that both of these policies required Qualcomm to license its SEPs to all applicants, including competing chipset manufacturers. Qualcomm argued that the IPR policies contain limitations and that Qualcomm is not required to license its SEPs to applicants, like chipset manufacturers, that only produce components of devices. The FTC applied for partial summary judgment on this point (the trial on the wider issue of whether Qualcomm’s actions have harmed competition is scheduled for January 2019).

The decision

Judge Koh preferred the FTC’s interpretation, taking into account the following points:

  • Non-discrimination: Judge Koh reasoned that “if a SEP holder could discriminate against modem chip suppliers, a SEP holder could embed its technology into a cellular standard and then prevent other modem chip suppliers from selling modem chips to cellular handset producers”. She suggested that such discrimination could enable a SEP holder to achieve a monopoly, in direct contradiction of the stated purpose of the TIA IPR policy.
  • Industry practice: as part of its argument, Qualcomm claimed that chipset manufacturers never receive SEP licences. However, Qualcomm itself had received licences to manufacture and sell components, and had received exhaustive licences from over 120 companies, indicating that it could not be contrary to industry practice for chipset manufacturers to obtain SEP licences. 
  • Prior litigation:  when defending a patent infringement case against Ericsson, Qualcomm had previously claimed that the TIA policy required Ericsson to license any patents ‘required to develop products compliant’ with a given standard. Qualcomm suggested that this requirement would enable all industry participants to develop, manufacture and sell compliant products, and importantly, that chipsets were ‘compliant’ products covered by the policy.
  • Implementing the standard: Judge Koh also dismissed Qualcomm’s arguments that chipset manufacturers do not practise its patents. She noted that neither the ATIS not TIA policy restricts a SEP holder’s FRAND obligations to applicants that themselves practice or implement a whole standard. She emphasised that Qualcomm’s own documents demonstrate that a modem chip is a core component of a cellular handset, and that such chipset implements key cellular technologies.

Judge Koh also referred to a Ninth Circuit precedent (Microsoft II) as establishing that that Qualcomm’s RAND commitments include an obligation to license to all comers, including competing chipset manufacturers. She noted that in Microsoft II the Ninth Circuit had been interpreting a SSO IPR policy with almost identical language to the TIA and ATIS IPR policies.

For all of those reasons, Judge Koh agreed that the test for partial summary judgment (that ‘the meaning of the contract is unambiguous’) was met, and that both IPR policies required Qualcomm to license its SEPs to chipset manufacturers.

Conclusion

Although it was only a hearing for partial summary judgment, this case aired a significant number of the issues and arguments involved in the ‘licensing to all’ debate. These were all dealt with thoroughly by Judge Koh, though Qualcomm is expected to appeal. 

Judge Koh’s reliance on Qualcomm’s previous conduct highlights a frequent problem for large SEP holders that act as both licensor and licensee; it is difficult for such companies to eliminate conflicting positions in different cases.  However, not every judge will place the same weight on such considerations: in Unwired Planet, Birss J dismissed the relevance of past statements made by Ericsson and other SEP holders as to the appropriate total royalty burden (although in TCL Judge Selna took the opposite approach).   

In any event, this ruling is unlikely to be the last word on the matter. The case primarily focussed on contractual interpretation, rather than on antitrust as such. Establishing a contractual requirement for Qualcomm to offer licences to competing chipset manufacturers is one thing. Whether Qualcomm and any chipset manufacturers can agree the terms of a (F)RAND licence is quite another, particularly given that the order relates only to the ATIS and TIA IPR policies and so has a direct impact only on licences to those of Qualcomm’s SEPs which have been declared to ATIS/TIA, rather than necessarily on its global portfolio.  It remains to be seen whether Qualcomm will elect to carry this finding across to SEPs declared to other standards bodies, such as ETSI, and whether chipset manufacturers will in fact benefit from an exhaustive licence from Qualcomm.

Unwired Planet v Huawei: Court of Appeal upholds Birss J’s judgment

The keenly awaited appeal judgment in Unwired Planet v Huawei was handed down yesterday. In its unanimous judgment, the Court of Appeal dismissed Huawei’s appeal, confirming Mr Justice Birss’s first instance decision (see previous commentary here and here) on FRAND licensing of standard-essential patents (‘SEPs’). We summarise the key findings below. 

1. Global licensing may be FRAND

The Court of Appeal held that a global licence can in principle be FRAND, and that if such a licence is refused by an implementer, the SEP holder should be entitled to the usual relief available for patent infringement, including an injunction. Agreeing with Birss J that Unwired Planet’s portfolio is substantial in size and scope and that Huawei’s business is global in nature, the Court decided that a global licence is what reasonable companies would have agreed on the facts of this case.

The Court of Appeal disagreed with Birss J on one point.  In the Court’s analysis, Birss J’s finding that there is only one set of FRAND licence terms in any given scenario was at odds with the complexities of patent licensing – and concepts such as fairness and reasonableness did not sit easily with Birss J’s rigid approach. This did not affect the Court’s overall conclusion on the global licensing issue, however. 

Implications

The Court of Appeal’s strong endorsement of Birss J’s approach means that the UK is likely to remain an attractive forum for SEP holders seeking to resolve global licensing disputes.  Nevertheless, the judgment still leaves scope for new issues to be raised: the Unwired Planet saga turned very much on its own facts and on Birss J’s assessment of those facts. The judgment does not mean that every future court-determined SEP licence will necessarily be global: this will depend on the nature of the portfolio to be licensed and the implementer’s sales and manufacturing footprint.  The law in this area is likely to continue to develop as new issues are raised (for example, an appeal of a jurisdiction challenge is currently pending before the Court of Appeal in another FRAND licensing case) and as other jurisdictions continue to develop their FRAND jurisprudence. 

2. Unwired Planet’s offer was non-discriminatory

The Court of Appeal agreed with Birss J that the ‘ND’ limb of the FRAND obligation only requires a SEP holder to offer a rate which reflects the proper valuation of the portfolio to all potential licensees. However, a SEP holder is not prevented from charging less than that benchmark rate if it chooses to do so. Indeed, the Court acknowledged that price discrimination is inherently neither pro- nor anti-competitive, stating that “an effects-based approach to non-discrimination is appropriate”. In cases where discrimination below the FRAND benchmark rate does cause competitive harm, this can be addressed by the application of competition law.

The approach of the Court of Appeal is in line with European competition law, which only identifies anti-competitive discrimination where there is a risk of competitive harm.  Arguably, the ruling at first instance went further, and required actual harm to be proved (the Court of Appeal did not address this point). 

Implications

SEP holders are likely to welcome this flexibility, which may also be beneficial for implementers who are looking to take advantage of a bespoke deal.  But in the context of litigation, implementers may be concerned that the ND obligation now lacks teeth. Nevertheless, prior licences of a particular portfolio are likely to remain of critical importance in any comparator licence analysis under the “fair and reasonable” limb of FRAND (unless they can be sufficiently differentiated, such that they are deemed not to be comparable), so the Court of Appeal’s judgment doesn’t give SEP owners a free hand to license at differential rates.

3. The Huawei v ZTE negotiation framework is not a set of prescriptive rules

The Court of Appeal has taken a pragmatic view, holding that just one part of the Huawei v ZTE framework is mandatory: the obligation on the SEP owner to contact and notify the implementer before starting litigation. The remainder of the framework is said to provide a ‘safe harbour’ – the licensor may stray from it, but in doing so faces risks of infringing Article 102 and being unable to obtain an injunction.

Implications

Whilst the Court has endorsed Birss J’s relatively flexible approach to applying the Huawei v ZTE framework, parties should nonetheless think carefully before straying too far from the scheme established by the CJEU.  Non-compliant conduct may remain high-risk if litigation occurs in the courts of other EU member states.

What next for FRAND?

The High Court’s judgment has been upheld and so the ‘FRAND injunction’ issued by Birss J will now take effect, unless a further suspension is ordered while an appeal to the Supreme Court takes place.  Huawei has already indicated that it may seek permission to appeal.

It remains to be seen to what extent the courts of other countries, including emerging FRAND centres such as China, will sit back and allow the UK courts to play ringmaster on FRAND/SEP issues. Recent guidelines issued by the courts of Guangdong (an important tech centre in China) suggest that global FRAND disputes may also find a home in other jurisdictions.

In any event, FRAND in the UK will continue to be influenced by developments in other countries, whether that is the approach of the US courts (as seen in cases like TCL v Ericsson), or the policies of the European Commission (which published a Communication on SEPs in November 2017), or guidance issued by other authorities such as the Japanese Patent Office (which issued a guide to SEP licensing negotiations in June this year).


SEP holders’ guidelines on IoT / 5G FRAND licensing

The Commission’s SEP Communication was designed to offer guidance on FRAND and SEP licensing. However, as we have noted before, it did not take a position on certain controversial issues such as use-based licences, or whether the FRAND obligation requires licences to be offered to any company that asks for one.  These issues will become increasingly important as the Internet of Things (IoT) continues to develop and more companies at different levels of the supply and distribution chain enter the SEP licensing arena for the first time.  

Earlier this year we reported on the establishment of two CEN-CENELEC workshops, one primarily backed by SEP holders, one primarily backed by implementers, that both seek to produce guidance on industry best practice for SEP licensing. We noted that each workshop was likely to take a different view on these kinds of issues. 

The first workshop, backed by Nokia and IP Europe, has now produced a first draft of its Guidance for licensing SEPs in 5G and the IoT. The draft is available for public review and comments until 13 December 2018. It sets out six principles that should apply in SEP licensing: 

  1. Owners of patent rights which are essential for using standardised technologies (SEPs) should allow access to that patented technology for implementing and using the standard. 
  2. Both the SEP owner and the potential licensee should act in good faith with respect to each other with the aim of concluding a FRAND licence agreement in a timely and efficient manner.
  3. Each party should provide to the other party, consistent with the protection of confidentiality, information that is reasonably necessary to enable the timely conclusion of a FRAND licence.
  4. “Fair and reasonable” compensation should be based upon the value of the patented standardised technology to its users.
  5. A SEP owner should not discriminate between similarly situated competitors.
  6. If the parties are unable to conclude a FRAND licence agreement within a reasonable timeframe they should seek to agree to third party determination of a FRAND licence either by a court or through binding arbitration. 

Taken at face value, these principles may sound uncontroversial.  However, implementers may feel that the devil is in the detail.  One point which is likely to provoke dissent is the question of who is entitled to benefit from SEP licences.  The document observes that there is usually one point or level in the supply chain where a SEP owner will choose to license its technology for a given product or service.  It is suggested that some consensus around this will simplify licensing, reduce costs for all parties and help maintain a level playing field between licensees. In other words, it indicates that ‘licensing to all’ is not required. 

The basis on which royalties are to be calculated is also likely to prove controversial.  In suggesting that compensation “should be based upon the value of the patented standardised technology to its users”, the document does not make clear whether this should include value attributable to the technology’s inclusion in a standard (the EC Communication and previous guidance in the Horizontal Guidelines suggests that such value should in principle not be included).  Other indicators which the Guidelines suggest may be considered include consumer demand, measurable benefits of the patented standardised technology, and the price difference between substantially identical products with and without the standardised technology.  This text suggests that there is indeed an intention to move away from the ‘incremental value’ rule, an issue which has recently been raised as part of the wider, significant debate about the role of antitrust in FRAND in the United States. (See for example, this letter by 77 former government enforcement officials and professors of law, economics, and business to Assistant Attorney General Delrahim, in which they criticise a number of speeches the AAG has made. They suggest that patent hold-up is a serious antitrust concern partly because “implementers are vulnerable to paying supra-competitive royalties based on the entire value of the product, not on the value of the patented technology”; the AAG’s response, supported by a number of other experts, is here.)

The second workshop is yet to publish the first draft of its alternative proposal. However, based on previous positions adopted by the Fair Standards Association (FSA) and ACT | The App Association, it seems likely that it will:

  • Stipulate that a patent owner cannot seek to increase the value of its patents by focusing on value created by downstream innovators and devices.
  • Call for an obligation on SEP holders to license only relevant patents – which may not necessarily be an SEP holder’s entire SEP portfolio.
  • Require licensing to any and all that seek a license.

In short, it is likely to take the opposing view to the first workshop on the key unresolved issues in the interpretation of FRAND.  

If the final products of each workshop are completely opposed to one another, they will not be particularly helpful as guidance to new potential licensees. It may require judicial or regulator intervention to resolve these issues before the full rollout of the IoT and 5G.  It remains to be seen where the first real determination of these kinds of questions will take place. One possibility is the FTC v Qualcomm case in California (discussed here), where the FTC is seeking a declaration that Qualcomm must be prepared to license competing chipset manufacturers. That could provide a persuasive authority on whether FRAND requires licences to be offered to all, although other jurisdictions may of course take a different view.

CLIP of the month: Strengthening Buyer Power as a Solution to Platform Market Power?

Each month we publish a ‘CLIP of the month’, a publication that we have found to be controversial or thought provoking (if you haven’t noticed this feature before, see above, look to the right of your screen, just below the header!).

This month’s CLIP (available here) comes from two of the CMA’s economists, writing on the dynamics of platform-to-business relationships, and the options for market solutions in the face of calls for greater regulation.  According to the authors, the key is finding a market-based counter-balance to the power held by platforms. The authors theorise that this could come from the collective bargaining power of the platform’s users. 

It is welcome to see individuals from within the competition authorities considering the alternatives to regulation, which could prove difficult in such a fast moving area.  It is also notable that the UK may seek to move in a different direction from many of the EU member states which appear to favour greater regulation.  However, such market-based alternatives may themselves face competition law challenges, given the risks that exist around collective bargaining and exchange of information between competitors.  At least some of these risks are recognised by the authors (and it is worth reading an economist’s perspective on this from David Parker of Frontier Economics, here).   However, without some kind of safe harbour for such discussion across all relevant territories, the competition law risks are likely to remain a significant disincentive to such collective action.  There may also be a problem of timing.  In the case of existing businesses, the incentive and ability to challenge platform market power may be at their highest before true platform market power emerges.  

The role and (potentially) regulation of platforms is likely to remain a key area of debate in competition policy for the foreseeable future.

Summer FRAND developments: some big antitrust news to come

Back in May this year, the Court of Appeal sat for five days to hear Huawei’s appeal of Birss J’s judgment in Unwired Planet.  The bench included two of the Court’s most experienced patents judges in Lord Justices Kitchin and Floyd (and Kitchin LJ has now been appointed to the Supreme Court, see here). Although rumours circulated that the Court of Appeal’s judgment might be forthcoming before the summer vacation, nothing materialised, and the expectation now is that the judgment will be published at the start of the Michaelmas term, which begins on 1 October. (A reminder of the issues being appealed can be found here.)

That judgment will be hugely significant for the conduct of FRAND negotiations, licensing and litigation both in the UK and elsewhere. This is not least because there are now several FRAND cases before the High Court, including Apple v Qualcomm, Conversant v ZTE & Huawei, Philips v ASUS & HTC and TQ Delta v Zyxel (the latter on ASDL, rather than ETSI telecoms standards). All of these are likely to be affected by the Court of Appeal’s decision.

However, the upcoming judgment isn’t the only relevant FRAND news. FRAND is a global concern, and recently there have been other FRAND developments around the world that are worth noting.

Delay to German Supreme Court ruling in Sisvel v Haier 

Back in 2015, the Düsseldorf Regional Court granted Sisvel an injunction against Haier for infringement of two its SEPs. However, on appeal, the OLG Düsseldorf held that Sisvel had failed to offer Haier FRAND terms as per Huawei v ZTE because Sisvel’s offer was discriminatory (there was a significant difference between Sisvel’s treatment of Haier and its competitors). Sisvel appealed to Germany’s Supreme Court, the Bundesgerichtshof.

In parallel, Haier had brought proceedings challenging the validity of the two Sisvel SEPs. One has now expired, and the other was declared invalid at first instance. Sisvel has appealed this decision.

The Supreme Court has now suspended its decision on the FRAND side of the case until the patent proceedings are completed, to avoid the risk of conflicting judgments (the decision is here, in German).  

It seems likely that if the invalidity decision is upheld, the Supreme Court will decline to make a final determination on the FRAND issues. Given the limited case law available on the discrimination aspect of FRAND (see here for some discussion of this), as well as the lack of consensus on the approach to FRAND in the German regional courts, this will mean that uncertainty continues at least as regards the German position for the foreseeable future.

Haier fails to convince US Court to take up digital TV case

Meanwhile, over in the USA, Haier had sought to turn the tables on other holders of declared SEPs, having brought a case in the Northern District of New York alleging a conspiracy by the holders of SEPs reading onto the ATSC standard.  Haier claimed that a patent pool established by companies including LG and Panasonic and administered by MPEG LA was artificially inflating prices above a FRAND level, and that licensors were refusing to license individually.

The Court rejected the claim on limitation grounds earlier this month (September 2018) and did not engage with the substance of the allegations (see here). 

FTC moves for summary judgment on Qualcomm’s obligation to license competitors 

In the Northern District of California, the FTC is engaged in proceedings against Qualcomm, alleging that Qualcomm has excluded competitors and harmed competition by withholding its baseband processors unless a customer accepts a licence on terms favourable to Qualcomm (including disproportionately high royalties). The trial is scheduled for early January 2019.

However, at the end of August, the FTC filed a motion for partial summary judgment regarding a key aspect of the case. It has asked for a declaration that under its FRAND obligations, Qualcomm must license its SEPs to its chipset competitors such as Intel.

Whether FRAND requires SEP holders to grant a licence to any company that asks for one (known as licensing to all) is a hotly debated topic.  The answer is potentially of wide significance, because it could fundamentally affect the licensing model that has applied in the sector for the past 20 years. For example, if all chipset manufacturers were licensed (potentially at royalties based on the chipset price rather than on the price of a smartphone) the manufacturers of smartphones may not require licences at all (depending on laws relating to pass-through and exhaustion) which would have a major impact on how SEP licensing currently operates.  Alternatively, the price of chipsets themselves might need to rise significantly to account for the increased IPR costs.  Or manufacturers may start to seek to tailor licences to different uses, splitting value along different parts of the supply and distribution chain. 

Examples of any judicial authority considering this topic are rare. The Commission dodged the issue in preparing its 2017 Communication on SEP Licensing (see here), although the Korea Fair Trade Commission has found Qualcomm’s refusal to license its SEPs to competing chipset manufacturers abusive (the decision is here, but note that Qualcomm is appealing). Although the California Court’s judgment will relate to the ATIS and TIA standards interpreted on the basis of US law (rather than ETSI FRAND interpreted on the basis of French law), it is still likely to be influential in Europe.

Big Data? No antitrust problem for Apple/Shazam

Big Data has been a focus for DG Competition for the last few years.  In particular, the Commission has been interested in mergers involving the acquisition of a company holding valuable data, even if it has low turnover (see here). Apple’s $400 million acquisition of Shazam, approved by the Commission on 6 September 2018, falls squarely within this category.

Shazam is a music recognition app.  Consumers can use Shazam to record a clip of an unknown song playing in a bar or other public place – Shazam turns this into an audio fingerprint and matches this against its database containing the audio fingerprints for millions of songs in order to identify the song playing.  Shazam generates revenue through referring users once a song has been identified to Spotify, iTunes, Google Play music or other streaming services.  Although Shazam has been downloaded over 1 billion times, and is used over 20 million times a day, in 2016 it made a loss of over £3.5 million on revenues of just over $40 million.  Apple isn’t buying Shazam for its profitability, but rather the data it possesses on its millions of users, including which songs they like to listen to and other trends.

There have been a string of mergers in which data has been an issue, starting with Google/DoubleClick in 2008 and including Facebook/WhatsApp in 2014, and Microsoft/LinkedIn in 2016.  In most of these cases, the data-related concerns have centred on potential barriers to entry arising due to the concentration of valuable data in the hands of one company. The Commission looked for example at whether other advertisers would be able to replicate the data held by Google after its acquisition of DoubleClick, or whether access to the LinkedIn database was an essential part of developing advanced customer relationship management technology using machine learning in the Microsoft case.  

In Apple/Shazam the concern was different. The Commission opened a phase II review of the transaction on 23 April 2018 (here) because it was concerned that access to Shazam’s data might enable Apple to target directly the customers of its rivals (such as Spotify or Google Play Music) to encourage them to switch to Apple Music.  Unlike the more general concern in previous cases that it might be harder for new players to enter the market, here the Commission was specifically concerned about the potential for active harm to Apple’s competitors.

However, following its review, the Commission concluded (here) that:

  • Access to Shazam's data would not materially increase Apple's ability to target music enthusiasts and any conduct aimed at making customers switch would only have a negligible impact. 
  • The merged entity would not be able to shut out competing providers of digital music streaming services by restricting access to the Shazam app; the app has a limited importance as an entry point to music streaming services. 
  • The integration of Shazam's and Apple's datasets on user data would not confer a unique advantage to the merged entity in the markets on which it operates; Shazam's data is not unique and Apple's competitors would still have the opportunity to access and use similar databases.
Commissioner Vestager released a short statement accompanying the clearance: 

"Data is key in the digital economy. We must therefore carefully review transactions which lead to the acquisition of important sets of data, including potentially commercially sensitive ones, to ensure they do not restrict competition. After thoroughly analysing Shazam's user and music data, we found that their acquisition by Apple would not reduce competition in the digital music streaming market."

There seemed to be a hint of disappointment that the Apple/Shazam transaction had not ultimately enabled the Commission to take real action against a big tech company hoovering up a smaller but data-rich target, but Vestager’s comments do reveal how seriously the Commission views the acquisition of data and its potential to pose a threat to fair competition. 

There are a number of other interesting aspects to this particular transaction. The Commission’s press release also explicitly stated that “a merger decision does not release companies from respecting all relevant data protection laws”.  We have previously looked at the privacy / competition overlap (here and here for example) and this is a pointed reminder from the Commission that whilst there are no competition concerns related to the data in this transaction, Apple will have to ensure that it treats the newly acquired data in a manner that complies with the GDPR.

Apple/Shazam did not meet the turnover thresholds set by the EUMR.  It was reviewed by the Commission only because Austria (the one Member State where the transaction did meet the national merger notification threshold) submitted a referral request to the Commission under Article 22(1) EUMR. This was the first time in over two years that a national competition authority had referred a transaction up to the Commission.  Austria’s request was joined by Iceland, Italy, France, Norway, Spain and Sweden – equal to the record highest number of authorities seeking a referral (here – subscription required).

The interest taken by these national authorities in this transaction reflects the increasing recognition of the importance of acquisitions of innovative companies. There have been some concerns that high value transactions of companies with low turnovers (but valuable data) may escape review by competition regulators. For example Intel’s 2017 $15 billion acquisition of Mobileye, an Israeli company manufacturing self-driving car technologies, avoided review in Europe due to Mobileye’s low revenues in Europe.  As we discussed here, Germany (like Austria) has recently amended its domestic competition law; changing its merger thresholds to try and capture these kinds of transactions. 

We suspect there will be plenty more phase II investigations of big data mergers in the future.

Online advertising – Government acknowledges challenges ahead for competition law

We reported a few months ago on the House of Lords Communications Select Committee's report on advertising in the digital age.

The Committee’s report set out the challenges currently facing the UK’s advertising industry in light of factors such as Brexit and the ever-expanding digital economy.  Notably, the Committee called on the CMA and other regulatory bodies to adopt robust standards for online advertising and made several recommendations to the Government of ways to help ensure that digital advertising “is working fairly for businesses and consumers”.

Following on from that report, the Government has now published its response.

What was the Government's response?

The following aspects will be of interest to competition lawyers:

  • The Government acknowledged that the regulatory challenges posed by the online advertising industry are extensive. The Government encourages continued self-regulation of online advertising, but may consider legislating in this area. A White Paper that focuses on online harms will be published later this year, which may shed light on any planned legislation relating to online advertising. 
  • In relation to the lack of transparency in the digital media advertising market, the Government stated that it is keen to gather more evidence on the business models in this market. This will hopefully form part of the Digital Charter's work programme and will be included in the Carincross review into the sustainability of the press (expected early 2019). However, it noted that because the CMA is an independent authority, the Government‘s powers to direct the CMA to undertake an investigation or study are extremely limited. 
  • Despite receiving an increase in funding, it remains unlikely that the CMA will have sufficient resources to fund a wide-ranging market study into digital advertising. This is because the additional £23.6 million funds allocated to the CMA by the Treasury will be going towards preparation for Brexit.  Market studies are cost intensive for the regulator, and many have speculated that they may be a likely casualty of the CMA’s increased enforcement responsibilities. 
  • The Government is to conduct an overall consumer markets review, to be completed by April 2019. As part of this review, the Government will consider how best to ensure the UK's competition framework is effective in responding to challenges presented by digital services.

The response to the Committee’s report indicates that the Government acknowledges the need to gather more evidence and to develop better tools in order to be able to deal with potentially complex competition issues arising in digital ad markets.  It also notes the overall preference for self-regulation (which should allow markets to self-correct), rather than introducing a rigid regulatory framework. However, with the CMA about to take on the burden of cases that would normally be dealt with by the European Commission, it remains to be seen whether the UK will be able to adapt quickly to the particular challenges posed by rapidly evolving digital markets. 

Commission considers pricing algorithms in fining consumer electronics manufacturers for RPM

Last year, we speculated whether the European Commission might target pricing algorithms (here) and noted that the Final Report of the Commission’s E-Commerce Sector Inquiry had identified the wide-scale use of pricing software as an issue that might raise competition concerns (here).

On 24 July 2018, the Commission announced that four consumer electronics manufacturers, including Asus and Pioneer, have been fined a total of over €111 million for imposing fixed or minimum resale prices on their online retailers (here). If the online retailers tried to set lower prices than those requested by the manufacturers, they were threatened with actions such as the withdrawal of supplies of the product in question.  

Interestingly, the Commission noted that the use of pricing algorithms by the online retailers exacerbated the impact of the manufacturers’ conduct. Many retailers were using pricing algorithms to automatically adapt their prices to those of their competitors. So where manufacturers were able to force some online retailers to adopt higher prices than they wished to, this had a broad impact across overall online prices.  It meant that other online retailers would match those higher prices, rather than the lower prices that might otherwise have been introduced.  In essence, this is rather akin to the umbrella price effects arising from many cartel-type agreements between competitors, but is not something that would – in the pre-pricing-algorithm era – have been so possible in the world of distribution agreements.

This isn’t an example of the Commission taking issue with the use of pricing algorithms in their own right. It hasn’t accused the online retailers of using pricing software as a means of coordinating on prices. Nor has it taken aim at the companies producing such software. The Commission did flag the manufacturers’ use of “sophisticated monitoring tools” to track resale prices, which enabled them to intervene quickly if a retailer attempted to decrease its prices, but otherwise pricing algorithms/software were part of the background to the decisions rather than the focus.

This may have been because there is nothing novel in the underlying competition law infringements committed in these cases; they appear to be classic cases of resale price maintenance contrary to Article 101 TFEU – although it is true that the Commission has not enforced against such agreements for a number of years before this announcement. In any event, there is nothing here to alarm companies making use of pricing algorithms or monitoring software, as long as they aren’t using such tools to implement an unlawful strategy.

However, these decisions do show that the Commission is alive to the potential for pricing algorithms and other software tools to be utilised as part of anti-competitive conduct, and such tools may feature more heavily in future Commission decisions. The Commission launched a number of other investigations following the E-Commerce Sector Inquiry (see here and here) into issues such as the licensing and distribution of merchandising products, the geo-blocking of PC video games, and hotel price discrimination on the basis of customer location.  Any decisions adopted following the conclusion of those investigations may offer some further guidance on the use of pricing algorithms.

MasterCard and Visa MIFfed as the Court of Appeal considers two-sided markets; SCOTUS itself is two-sided (Part 2 – the USA)

Following on from yesterday’s blog on the MasterCard / Visa decision, we’ve also taken a look at how the US is approaching antitrust issues in two-sided markets, with SCOTUS giving its first Opinion on these in the AmEx litigation (originally brought with the DoJ, but continued by only eleven states following the administration change). 

AmEx is a closed loop network, with AmEx holding relationships with Cardholders and Merchants.  In a 5-4 decision considering anti-steering provisions that prohibited merchants from avoiding fees by discouraging AmEx use at the point of sale, SCOTUS found no violation of the Sherman Antitrust Act (upholding the U.S. 2nd Cir. Court of Appeals).  SCOTUS was asked to determine whether the parties had met the burdens in a three step rule of reason analysis (plaintiff must prove anticompetitive effects; defendant must show a procompetitive justification; plaintiff must show that efficiencies could have been achieved through less restrictive means).  

The plaintiffs sought to argue that a market definition wasn’t necessary because they had offered evidence that showed adverse effects on competition.  The majority disagreed with this, and noted that the cases relied on by the plaintiffs for this proposition were horizontal restraint cases. Here, vertical arrangements were at issue, and given that they’re not always anti-competitive, the market definition was relevant.  

The majority held that the relevant market included both sides of the transaction, and it had to be shown that both sides of the transaction were harmed by the provision.  The Court stated that “a market should be treated as one sided when the impacts of indirect network effects and relative pricing in that market are minor”, and divided two-sided markets into two categories:

  1. Two-sided transaction platforms that facilitate a single, simultaneous transaction and are best understood as supplying that transaction as the product (such as those between AmEx Cardholders and Merchants); and

  2. Platforms where the two sides aren’t selling directly to each other (such as newspapers, where users are indifferent to the amount of advertising).

Further, with the first category, evaluating both sides would be necessary to assess competition; only other two-sided platforms can compete for transactions.  Non-transaction platforms often do compete with companies that do not operate on both sides.  Unfortunately for the plaintiffs, their evidence was insufficient as they had only focused on the increase to merchant fees.  This division will perhaps create some debate as to which category a platform falls into, and arguments around how strong indirect network effects are. 

The majority stated that in order to show that the provisions were anticompetitive, plaintiffs should have demonstrated that they increased the cost of credit card transactions above a competitive level, reduced the number of transactions, or otherwise stifled competition.  In fact, the majority found that the provisions were pro-competitive, as they helped maintain a competitor to MasterCard and Visa. 

The dissenting opinion, which included some persuasive points, wanted the US to follow other jurisdictions and take action against high fees charged by credit-card companies to Merchants, viewing the provisions as clearly anticompetitive.  It referred to findings by the District Court, and stated that a market definition was unnecessary because of direct evidence of anticompetitive effects (primarily that AmEx was able to keep increasing fees without losing any large Merchants), but the correct relevant market should have been only the one side – the services are complementary, not substitutes – and that the other side of the market should have come in at the second step of the rule of reason analysis.  This perhaps puts the dissenting justices more in line with the CoA’s approach.   

Dissenting, Justice Breyer further countered the view that the provisions were pro-competitive by stating that “if American Express’ merchant fees are so high that merchants successfully induce their customers to use other cards, American Express can remedy that problem by lowering those fees or by spending more on cardholder rewards so that cardholders decline such requests”.  

Back in the UK, the Merricks collective claim is attempting to show that harm was caused to consumers –not on the flip side of the market, but by Merchants passing on the cost of the MIFs to customers.  Although the CAT refused to allow the action, the appeal is due to be heard later this year.  Whilst there is quite a hurdle to jump in how to ensure consumers receive compensatory amounts rather than token sums of money, if the class is certified, the analysis of effects on consumers and the links between the different markets could make for interesting reading. 

(On a collective action side note…  After two years of build-up, it’s good to see that the first trucks collective claim has started rolling towards certification, and interestingly, is using a special purpose vehicle more typical of litigation in the Netherlands than the UK.)

MasterCard and Visa MIFfed as the Court of Appeal considers two-sided markets; SCOTUS itself is two-sided (Part 1 – the UK)

Whilst the Court of Appeal’s judgment in MasterCard / Visa, and the SCOTUS Opinion in AmEx may seem a little outside our usual area of focus, they are nevertheless decisions that relate to the operation of two-sided markets.  With multi-sided platforms in innovative technological markets, such as Google, Facebook and Uber, increasingly drawing antitrust attention, (see here, and here) there may be some helpful guidance to be drawn from long established industries such as banking and finance. 

This post comes in two parts, with today focusing on the MasterCard / Visa judgment, and tomorrow focusing on the AmEx litigation. 

The Court of Appeal judgment 

Both MasterCard and Visa operate four-party payment schemes:

  • Cardholders contract with an Issuer for a card to buy goods from Merchants;
  • Merchants contract with Acquirers to obtain payment from the Cardholders;
  • Issuers (mostly banks) contract with Acquirers (also mostly banks) to settle transactions.

The Issuers compete for the business of the Cardholders, and the Acquirers compete for the business of the Merchants; but each side is dependent on the other.  The MasterCard / Visa schemes operate as open loop networks, and those participating are subject to various rules – including a requirement to pay fees, including multi-lateral interchange fees (‘MIF’s), that are charged by the Issuer to the Acquirer, and ultimately paid by Merchants in each card transaction.  The MIFs could have been negotiated individually between the Issuer and the Acquirer, but in practice default MIFs set by MasterCard / Visa were used.   

This raised an interesting Article 101(1) question: do the schemes’ default MIFs amount to a restriction of competition by effect?  The European Commission thought so in issuing a 2007 decision against MasterCard in respect of cross-border card transactions, a decision which spawned a multitude of follow-on and standalone actions for damages against both MasterCard and (by analogy given the similarities between their systems) and VISA.  The CAT initially found for the Claimant in one damages action, but the High Court subsequently found for the Defendants in separate actions (MasterCard and Visa).  The Court of Appeal was tasked with addressing these inconsistent outcomes.

The systems themselves operate across three separate markets (an inter-systems market, an issuing market, and an acquiring market), and it was common ground that the relevant market was the acquiring market.  However, arguments raised by the parties (particularly the ‘death spiral’ argument, where MasterCard claimed that if it lowered its MIFs, Issuers would have switched to Visa and the MasterCard scheme would have collapsed) concerned effects on the inter-system market, and the issuing market.  The CoA held that the first question is whether the MIFs restricted competition in the acquiring market. The second question is then whether the MIFs were objectively justified, and at that point, it is legitimate to consider both sides of the two-sided market and the inter-system market. 

The CoA ultimately found that the fees were unlawful, and all three cases are to be remitted to the CAT for an assessment of damages, and a determination as to whether any objective justification applies.  Tomorrow, we’ll set out how the US Supreme Court came to the conclusion that provisions which affected Merchants’ transaction costs were not anti-competitive, with analysis turning on a definition of the market that has implications for all platforms.