Daniel S Bitton and Leslie C Overton, Axinn, Veltrop & Harkrider LLP

This is an extract from the second edition of the E-Commerce Competition Enforcement Guide - published by Global Competition Review. The whole publication is available here.

When data is the product

The US agencies have on multiple occasions intervened in mergers involving commercially valuable data, but primarily when that data was the actual competitive overlap product.

Examples of such cases include the FTC’s 2014 challenge of Corelogic’s proposed acquisition of DataQuick; its 2013 challenge of Nielsen’s proposed acquisition of Arbitron; its 2010 challenge of Dun & Bradstreet’s already consummated US$29 million acquisition of Quality Education Data (QED); its 2009 challenge of Reed Elsevier’s US$4.1 billion proposed acquisition of ChoicePoint; and the DOJ’s 2008 challenge of Thomson’s proposed US$17 billion acquisition of Reuters. In each of those cases, the merging parties were the top competitors in highly concentrated markets for a specific data service.

Notably, however, none of these deals involved e-commerce, or even really big data. Rather, each of these deals involved more traditional business-to-business database services involving data that was difficult and costly to replicate for remaining rivals and new entrants because of the extensive and often manual effort required to collect the data.

For example, the relevant product in Corelogic/DataQuick was a database of ‘national assessor and recorder bulk data’. Collection of that data involved a rather manual process of extracting public information (such as ‘parties to the transaction, transfer tax, and purchase price’) from transactional documents like deeds, mortgages, liens, assignments and foreclosures available through local government offices. The relevant product in Nielsen/Arbitron was a database of audience measurement data. That was difficult for future rivals to replicate because only the parties had the capability to measure consumers’ radio and television consumption on a national level given the survey panels that they had put together over many years. The relevant product in Dun & Bradstreet/QED was a database of ‘contact, demographic and other information’ relating to Kindergarten through twelfth grade teachers, administrators, schools and school districts used for ‘direct marketing mail and email marketing efforts’. Matching the merged entity’s database involved a rather manual effort of collecting and verifying schools’ and educators’ contact information for many schools and teachers across the country. The relevant product in Reed Elsevier/ChoicePoint was databases of a ‘wide array of public and non-public records about individuals and businesses, including credit header data, criminal records, motor vehicle records, property records, and employment records’. The parties ‘compile[d] these records, either by going directly to the source or by purchasing these records from third parties’. The relevant products in Thomson/Reuters were three types of financial databases, that were difficult to replicate, among other reasons, because it involved: ‘harvesting’ and ‘normalizing’ information from the financial statements of ‘thousands of companies’ spanning ‘many years’; ‘obtaining the research report from a wide range of brokerage houses and other financial institutions’ in some cases ‘going back years or decades’; and ‘aggregat[ing]’ research reports from ‘hundreds of investment banks and brokerages’. 

In contrast to the types of discrete but hard-to-collect data sets in these cases, the challenge with big data typically is not that it is difficult or costly to collect. Quite the opposite. Owing to the widespread and prolific consumer use of countless online and offline services and the increasingly low cost of computing power (in part due to cloud services), the challenge with big data is typically that it is so voluminous, fast-growing and ubiquitous that many companies collect it in spades but struggle with what to do with and make of it. As reported by several data analytics companies, it is the advanced technologies that extract actionable information from large data sets in which the key differentiator and value lies in big data. 

Cases like Corelogic/DataQuick thus suggest that the agencies typically have had fewer concerns about two companies merging big user data sets than two parties merging specialised, manually collected (and therefore difficult to replicate) databases. This was reinforced during the recent FTC Hearings where the Acting Deputy Director of FTC’s Bureau of Competition, Haidee Schwartz, discussed the particular features of the data in that case: ‘So it is public information, but it is not standardized, it is not easy to collect, and you need both historical and going forward.’ Deputy Director Schwartz particularly highlighted the complexities involved in attempting to remedy the situation by positioning another player to be an effective competitor:

[T]he breadth, detail, and the complexity of the data created barriers to entry . . . You would think it is a database, it is not that hard to transfer, but here, the buyer’s due diligence may not necessarily uncover missing or unnecessary data in a timely fashion, and the Commission had difficulty initially identifying the exact universe of data required to effectively compete and required additional work by the buyer, the monitor, and the Commission to determine what data was missing, how it needed to be delivered, and how it needed to be continuously updated. 

Big data as an essential input and entry barrier

The agencies have also evaluated the theories of harm on big data currently in vogue. So far, however, those theories of harm have not led to merger challenges or remedies.

For example, when Google looked to acquire DoubleClick (a digital ad serving business) in 2007, several competitors apparently complained, among other things, that ‘the combination of Google’s database of user information and the data respecting users and competitive intermediaries collected by DoubleClick on behalf of its customers would give Google an overwhelming advantage in the ad intermediation market’. The FTC rejected that and other arguments against the acquisition. It closed its investigation of this transaction without any remedies, concluding (among other things) that there was no support for the proposition that the combination of these data sets would give Google market power, and that:

At bottom, the concerns raised by Google’s competitors regarding the integration of these two data sets – should privacy concerns not prevent such integration – really amount to a fear that the transaction will lead to Google offering a superior product to its customers. Yet, the evidence indicates that neither the data available to Google, nor the data available to DoubleClick, constitutes an essential input to a successful online advertising product. A number of Google’s competitors have at their disposal valuable stores of data not available to Google. 

Big data complaints seemingly have continued to be unsuccessful in US merger reviews since Google/DoubleClick.

For example, some raised similar concerns in Facebook’s 2014 acquisition of WhatsApp, and the European Commission evaluated (though rejected) data monopoly theories in that deal. But the FTC cleared the transaction unconditionally on antitrust grounds within two months.

When Microsoft won the bid to acquire LinkedIn in 2016, Salesforce.com reportedly expressed concerns to US agencies and the European Commission that:

by gaining ownership of LinkedIn’s unique dataset of over 450 million professionals in more than 200 countries, Microsoft will be able to deny competitors access to that data, and in doing so obtain an unfair competitive advantage. 

This concern was based on a more traditional vertical foreclosure theory, and in part also featured more traditional ‘data as the product’ theories like the ones at issue in Dunn & Bradstreet/QED, since LinkedIn did in fact sell its data as part of a sales intelligence solution to customer relationship management software providers like Salesforce. However, the theory of harm was similar to the one advanced in Google/DoubleClick, in that LinkedIn’s data was alleged to be an essential input for rivals. The FTC cleared the transaction without any remedies, apparently declining to challenge the deal on this basis. The European Commission did impose conditions in the transaction but not related to the data aggregation theory.

Third parties apparently also raised concerns about data aggregation (such as the ones raised in Google/DoubleClick and Facebook/WhatsApp) in the FTC’s 2017 investigation of Amazon’s acquisition of Whole Foods. But the FTC (under President Donald Trump) cleared the transaction without a second request. In discussing this transaction at a recent event, the FTC’s Bureau of Competition Director addressed another emerging and popular big data antitrust theory – that users sharing their data with online service providers is a form of paying for those providers’ services when they are otherwise free and that therefore aggregation of user data through a merger could have an effect equating to a price increase if it leads users to share more data with the merged company. He explained that there are several challenges with such concepts and theories of harm. One of the complicating factors is that not every consumer values his or her data the same way or at all (in economic terms). Another is that users sharing their data with online services often directly leads to the improvement of the service that they consume. Accordingly, the aggregation of user data through a merger, by enabling improvement of user services, could well lead to a price decrease – rather than increase – on a quality-adjusted basis (i.e., the quality of the service increases while the price does not). 

More recently, the European Commission launched an in-depth investigation of Apple’s proposed US$400 million acquisition of Shazam, on the theory that Shazam’s trove of data would give Apple an unfair competitive advantage over its music streaming rivals by enabling Apple to target users of rival services to encourage them to switch to Apple. In contrast, no such investigation or concerns appeared to exist at the US agencies insofar as we can tell. Even the European Commission eventually cleared the Apple/Shazam deal unconditionally. 

Thus, while the European Commission has seemingly entertained theories of harm based on big data concepts to a greater degree than the US agencies, we are not aware of either the European Commission or US antitrust agencies intervening in mergers based on such theories. That is not surprising because, even if there were merit to such theories in a particular case, tailored remedies for such theories – typically some sort of forced sharing of user data with rivals – would often be fraught with peril given the privacy implications.

Such remedies would also likely be unappealing to Trump administration antitrust officials for yet another reason: because they would force sharing of intellectual property between rivals, something that they are not naturally incentivised to do, and thrust enforcement agencies and courts into a regulatory role. The Trump administration officials at the FTC and DOJ have expressly said they will always look to avoid such remedies, if at all possible. That policy position, combined with agency officials’ recent scepticism about big data theories and the apparent rejection of such theories in Amazon/Whole Foods, would seem to suggest that this topic is unlikely to rank high on the merger enforcement agenda for the Trump administration.

Recently, the Assistant Attorney General of the DOJ made some of these points in a speech, among other things expressing reservations about the proposition that large user data sets are likely to create entry barriers or market power; highlighting that use and aggregation of such data in fact often significantly improves online services and advertising; and that forced data-sharing remedies can lead to undesirable policy outcomes. 

Data privacy considerations

From time to time the agencies have also been asked to include privacy considerations in the antitrust analysis of mergers involving big data. While US agency officials have acknowledged that privacy conceptually could be one quality parameter on which companies compete, they have generally rejected, both in speeches and in matters like Google/DoubleClick, the suggestion that antitrust merger review should be used to protect user privacy. That said, the FTC’s Bureau of Consumer Protection has shown it is willing to raise consumer protection concerns about privacy in the merger context, when merited. It did so, for example, when the FTC cleared Facebook’s acquisition of WhatsApp, by sending Facebook a letter reminding it to abide by WhatsApp’s privacy commitments to users. However, at least two of the FTC Commissioners have made clear their views that data privacy considerations when rooted in consumer protection concerns alone have no place in merger review. For example, Commissioner Christine Wilson recently stated, ‘If firms compete on the basis of privacy or data policies to attract customers, we might properly consider those aspects of non-price competition. But if firms do not compete that way, then they are appropriately omitted from our competition assessment.’ At the FTC Hearings, Commissioner Noah Phillips voiced concerns over enforcement from the FTC with respect to privacy considerations during antitrust review and stated that the proper balance between competition and privacy goals should be left to Congress. Former Acting Chairman Maureen Ohlhausen similarly provided that she has ‘concerns . . . with proposals to use antitrust to stop mergers or acquisitions by data-rich companies simply to address privacy concerns, not where the transaction or the behavior reduces privacy as a nonprice attribute of competition’.

Nascent competition theories

Some have suggested that the agencies should use the potential competition doctrine more aggressively to prevent incumbent online companies from acquiring tech start-ups that might challenge them in the future. They have cited the FTC’s challenge of Questcor’s acquisition of the rights to Synacthen from Novartis as a model for such an approach. There certainly are scenarios conceivable when that would be justified – for example, if the acquirer is an incumbent with market power and the start-up forms a unique competitive threat to that market power, unlikely to be replicated by others. Though, unlike in Questcor/Novartis, which apparently involved pharmaceuticals that were very difficult to develop and commercialise, establishing that a tech start-up did something unique and not replicable may prove challenging. The agencies have already investigated start-up acquisitions on that basis, such as Facebook’s 2014 acquisition of WhatsApp, and prior to that, in 2012, Instagram. In each case, however, both the US and EU agencies that investigated such transactions ultimately cleared them without conditions.

In Facebook/WhatsApp, the European Commission did so because the parties’ messenger apps were not close substitutes, faced several other messenger apps, and barriers to entry and expansion were not significant since users use multiple messenger apps at the same time and thus switch easily, such that anticompetitive effects were unlikely. In Facebook/Instagram, the UK Office of Fair Trading (OFT) concluded that the transaction did not raise concerns because there were more significant photo app competitors to Instagram than Facebook at that time, and that Instagram in turn was not a likely significant future competitor to Facebook in online advertising, in which Facebook faced more significant players. 

While the FTC’s considerations for clearing these were not public, they presumably reached the same conclusions as the European Commission and OFT. Notably, the US agencies have not been skittish about taking enforcement action based on the potential competition doctrine either. They did so recently, for example, in Nielsen/Arbitron and Steris/Synergy, and even litigated (though lost) in the latter matter. And, unlike in the incumbent/start-up scenario, in both of those cases, the merging parties were not even competing with each other yet with a live product.

In the start-up acquisitions, however, it seems that the agencies simply concluded that the facts at hand did not support a potential competition concern. In reaching those conclusions, the agencies undoubtedly considered the speculative nature of potential competition predictions in dynamic and fast-moving digital markets, in which the success of a start-up itself on the one hand suggests relatively low barriers to entry, while on the other hand is no guarantee that it (rather than another, more established player or new entrant) will become a significant or unique rival to an incumbent. But, as the FTC’s challenge of the Questcor/Novartis transaction shows, the nascent nature of markets certainly has not deterred the agencies from closely investigating and challenging mergers in the past. The FTC is currently reviewing a similar scenario with its investigation of Illumina’s acquisition of PacBio. According to the FTC, Illumina is the dominant provider of short-read gene sequencers, and PacBio of long-read gene sequencers, a separate market. The FTC is reportedly concerned that, because long-read gene sequencers could drop in price, the two markets could converge, making PacBio a nascent competitor. The FTC issued a second request on 2 January 2019. The deal is under Phase 2 review by the Competition and Markets Authority, with a decision due by 11 December 2019. 

Groups such as the New Brandeisians believe that previous administrations did not go far enough in using the potential competition doctrine to investigate and challenge start-up acquisitions, and the FTC’s current Chairman has indicated that the agency will spend more resources on acquisitions of nascent competitors by digital platforms, noting that the ‘harm to competition’ from such acquisitions ‘can . . . be significant’. Indeed, FTC Chairman Simons stated that a focus of the new Technology Task Force will be ‘monopoly maintenance issues’, specifically through the acquisition of nascent competitors. Commissioner Rebecca Slaughter and former Acting Chairman Maureen Ohlhausen confirmed their views that the creation of this task force increases the likelihood of enforcement. And, the Commission dedicated two panels and three presentations on nascent competition during the recent FTC Hearings, signalling to many the importance that the Commission and Chairman Simons specifically place on this issue. But it is not yet clear that the Trump administration will ultimately enforce more aggressively on this particular issue. While the FTC will apparently spend more resources investigating such acquisitions, the FTC Chairman has also acknowledged that acquisitions of nascent competitors in the high-tech space are ‘particularly difficult for antitrust enforcers to deal with because the acquired firm is by definition not a full-fledged competitor’ and ‘the likely level of competition with the acquiring firm is frequently, maybe more than frequently, not apparent’. The agencies may be cautious about litigating cases in such speculative circumstances. Moreover, this administration has shown sensitivity to preserving investment incentives (e.g., in IP) and therefore may be also hesitant to intervene in mergers based on relatively speculative theories of harm as a matter of policy – after all, for many entrepreneurs the prospect of being acquired by a large tech company is one important incentive to invest in a start-up in the first place. The following statement by the DOJ’s Assistant Attorney General for Antitrust is instructive in this respect (though it ostensibly was not about mergers):

We must remember that big platforms were once themselves start-ups, and be cautious in any enforcement decision to not undermine the very innovation incentives that competition aims to protect. At the Antitrust Division, we continue to study this area and believe in a careful application of the antitrust laws that takes into account both the short-term and long-term effects on innovation. We should take action only with credible evidence of harm to competition and not harm to just competitors. We must balance the goals to protect the very incentives to innovate, but at the same time be prepared to intervene when anticompetitive conduct distorts the free market. 

Traditional horizontal theories

The US agencies, in recent years, have also investigated several mergers in the digital space based on straightforward horizontal theories of harm, such as unilateral effects.

For example, in 2015, after an extensive investigation, the FTC cleared unconditionally Zillow’s US$3.5 billion acquisition of Trulia, the first and second largest consumer-facing online portals for buying homes. The parties’ internal documents suggested that they competed head-to-head to offer users home sale information and sell advertising to estate agents. The FTC nevertheless cleared the transaction without remedies because the data showed that the companies represented ‘only a small portion of agents’ overall spend on advertising’, and that their portals did not generate a higher return on investment for agents than did other forms of advertising used by the agents. This implied that the parties could not realistically increase prices post-merger without losing too much agent advertising spend to other forms of advertising. The FTC also found that the companies competed with a number of other portals to offer home buyers relevant information.

This case illustrates an important point to remember in mergers between online advertising businesses. Even if the merging parties attract consumers with similar online content, they often compete with a much broader array of (online) companies in selling advertising, since the same consumers typically can be targeted through many different advertising media. Zillow/Trulia is also a good reminder always to look closely at the parties’ data, because it may prove to be an important reality check on documents that paint an unhelpful but inaccurate or incomplete picture.

The same year DOJ investigated and cleared without remedies another major consolidation of online names: online travel agency Expedia’s US$1.3 billion acquisition of rival Orbitz (after Expedia had earlier acquired Travelocity). DOJ cleared the transaction despite vigorous complaints by hotels that they would have to pay higher prices post-merger to feature on Expedia or Orbitz. DOJ found that: there was no evidence suggesting that consumers would be charged higher fees for using the websites; airlines, hotels and car rental agencies barely received any bookings from Orbitz anymore, such that Orbitz did not exercise a significant constraint on Expedia’s commission charges – Priceline was by far Expedia’s most significant rival; and there were dynamic changes occurring in the industry with ‘the introduction of TripAdvisor’s Instant Booking service and Google’s Hotel and Flight Finder’. This is an example of a case in which the dynamic nature of digital markets and recent entry developments can be a legitimate reason for agencies not to intervene.

That said, the DOJ’s successful challenge of Bazaarvoice’s consummated acquisition of PowerReviews the previous year shows that a merger defence that online markets are dynamic only goes so far and that unhelpful documents still can kill deals. Bazaarvoice’s documents showed that its intent behind the acquisition was to eliminate its closest competitor in the sale of ‘product ratings and review platforms’. Following trial, the District Court found for DOJ, concluding that ‘[g]iven the overwhelming market share Bazaarvoice acquired when it purchased PowerReviews, the stark premerger evidence of anticompetitive intent and the merger’s likely effects, coupled with the actual lack of impact competitors have made since the merger, the government established the Section 7 violation’. Bazaarvoice was ordered to divest the PowerReviews business, notably in a way that re-established PowerReviews as if it had never been acquired (taking into account how it would have developed on its own but for the acquisition). 

Two other, more recent, horizontal deals between head-to-head competitors in the online space likewise did not fare well. In 2017, the FTC challenged the combination of Red Ventures’ A Place For Mom and Bankrate’s Caring.com, and entered a consent decree under which the merging parties agreed to divest one of those two assets (the merging parties’ portfolios extended beyond these assets). The overlapping businesses were alleged to be each other’s closest competitors in the provision of third-party paid referral services for senior living facilities (referral services ‘provid[ing] leads of qualified consumers to . . . senior living facilities’). Earlier that year, the FTC also had sued to block the merger of DraftKings and FanDuel, the two dominant online platforms for daily fantasy sports; a merger that it claimed would have resulted in a ‘near monopoly’. The parties competed vigorously on a number of elements, including: commission rates; discounts; contest prizes; and non-price factors, such as contest size, product features and contest offerings. Despite the unique and relatively nascent nature of this industry, the FTC pursued a fairly traditional unilateral effects case of closeness of competition. The parties ultimately abandoned the deal approximately a month after the FTC’s complaint. 

The latter two cases were both brought under the Trump administration, albeit before the new commissioners were put in place. It shows that this administration, like its predecessor, continues to enforce vigorously, including in online markets, based on traditional theories of harm.

In the area of horizontal deals, the current Chairman of the FTC and the Bureau of Competition director have indicated that they are both focused on stricter requirements for divestiture remedies. This is an area where the Trump administration could prove ‘tougher’ than its predecessor. A retrospective study of merger remedies done under the Obama administration found that while 80 per cent of all divestiture remedies were effective, only 70 per cent of those divestiture remedies that involved asset divestitures (as opposed to standalone business divestitures) were successful. As part of his Senate confirmation statement, Chairman Simons indicated that he plans to reduce that failure rate, which could suggest that the FTC will become even less receptive to asset divestitures (and more insistent on full business divestitures) than before. Of course, in practice, the scope of the business that needs to be divested will depend significantly on the identity and capabilities of the divestiture buyer. But the apparently tougher stance of this administration on divestiture remedies is something to keep in mind when consolidating major online competitors.

Non-horizontal theories

There have been a number of major non-horizontal merger reviews involving digital market companies in recent years, such as Apple’s acquisition of Shazam, Amazon’s acquisition of Whole Foods, Microsoft’s acquisitions of Skype and LinkedIn, and Google’s acquisition of ITA Software. Except for the latter, the US agencies did not conduct in-depth investigations of these transactions, much less seek remedies. In contrast, the European Commission subjected most of the same transactions to extensive investigation based on non-horizontal theories of harm, and even imposed behavioural remedies in one that the US agencies cleared without a second request.

For example, Apple’s acquisition of Shazam triggered an in-depth investigation from the European Commission (though it was cleared unconditionally) based on concerns that Apple would use Shazam data to target customers of rival music streaming apps (e.g., Spotify), as discussed above, and foreclose such rival apps from important referral traffic from Shazam. In contrast, the US agencies do not appear to have conducted such an in-depth investigation of the deal.

Similarly, the European Commission investigated foreclosure theories in Microsoft’s acquisitions of Skype and, especially, LinkedIn. In the Skype acquisition, the European Commission evaluated whether, after the acquisition, Microsoft would degrade Skype’s performance with other operating systems and platforms (or alternatively degrade how other communications services work on Microsoft’s Windows OS); integrate Skype with Windows or its Office productivity software, thereby reinforcing its dominant position as a ‘must-have’; or bundle the products to have the same effect. The European Commission cleared the transaction without remedies, concluding that while Microsoft could do those things, it was unlikely to do so since it would harm the Skype brand and drive users to rival communication services. In the LinkedIn acquisition, the European Commission tested a similar foreclosure theory (in addition to the foreclosure theories based on the LinkedIn data discussed above). Yet it came to the opposite conclusion and imposed remedies. The European Commission was concerned that Microsoft would pre-install LinkedIn on all Windows PCs and integrate it into Microsoft Office (among other applications), thereby significantly enhancing LinkedIn on dominant platforms such as MS Windows and Office, at the expense of rival professional social networks. As remedies, it required Microsoft to permit PC manufacturers and distributors not to install LinkedIn on Windows; to let users remove LinkedIn from Windows; and to allow rival professional social networks to maintain interoperability with MS Office and access to Microsoft Graph (to enable access to data stored on the Microsoft cloud). 

The US agencies, on the other hand, cleared the Skype and LinkedIn acquisitions without a second request or remedies.

The FTC likewise passed on a host of non-horizontal theories of harm put forth by opponents in its review of Amazon’s acquisition of Whole Foods. In addition to big data theories (discussed above), critics expressed concern, for example, that Amazon’s acquisition of Whole Foods would allow it to leverage its scale, logistics and buyer power in other retail areas to quickly dominate the grocery business (just like it did with book retailing). They also raised the concern that Amazon would be able to squeeze certain food suppliers with that dominance. The FTC cleared the acquisition promptly, without a second request, rejecting these types of conglomerate monopoly leveraging theories apparently for lack of cognisable antitrust harms. It did so presumably in part since both Amazon and Whole Foods had modest footprints in the online and offline grocery retail business. 

Of course, this is not to say that the US agencies will give any non-horizontal merger in the online space a pass. While the US agencies seem more sceptical of non-horizontal theories of harm in mergers involving online markets than their European counterparts, they have intervened in such matters in the past. For example, under Obama, the DOJ sought behavioural commitments to clear Google’s acquisition of airfare pricing and shopping software developer ITA Software. The remedies were designed to ensure that Google would continue to provide rival online travel websites such as Bing and Kayak access to ITA Software’s airfare pricing and shopping engine to power their flight search. Under Trump, the DOJ recently challenged – to the point of full litigation through trial and appeal – AT&T’s acquisition of Time Warner, which also was a purely vertical merger. While that acquisition was not entirely in the online markets sphere, the rationale for the transaction as well as its likely effects are very much so. AT&T, after all, claimed it pursued the transaction to compete directly in advertising with online advertising companies such as Google and Facebook. The DOJ sought to block the AT&T/Time Warner deal, out of concern that once part of AT&T, Time Warner would extract higher rents for its marquee programming (e.g., CNN, HBO) from traditional and online video distribution rivals of AT&T, such as Cox, Dish or Dish Sling, and thereby weaken their ability to put competitive pressure on AT&T. The DOJ lost its challenge both at the district court and appellate court levels, and the merger will stand. 

However, the outcome of the AT&T/Time Warner case could cause the US agencies to become yet more selective in enforcing against vertical mergers than they have been from a policy perspective, especially in dynamic online environments. Under Trump, the agencies have expressed strong reservations about behavioural remedies and indicated that they will typically insist on divestiture remedies if they have serious concerns, even in the rare vertical merger challenge (as DOJ did in AT&T/Time Warner). The outcome in that case shows, however, that there is a greater risk associated with taking that position in vertical deals (as opposed to accepting behavioural remedies, as done in the past), if it leads to litigation. Accordingly, the US agencies may be inclined to challenge fewer vertical deals, unless they are willing to revisit to some degree their policy position on use of behavioural remedies.

Multisided market definition

While an in-depth discussion of the Supreme Court’s recent decision in Ohio v. American Express is beyond the scope of this contribution, a few quick observations about that decision are worth making, given how frequently online markets and big data issues involve multisided platforms. The Supreme Court concluded that in proving the relevant market and market harm in cases involving multisided transactional platforms, such as American Express’s credit card network, plaintiffs must account for the effects of the conduct at issue on all customers of the platform (not just the effects on customers on one side of the platform). The court made clear, however, that this requirement did not uniformly apply to all multisided platforms, but rather only to ‘transaction’ platforms that ‘cannot make a sale to one side of the platform without simultaneously making a sale to the other’, explaining that in such a case a platform is ‘better understood as “supplying only one product” – transactions’. The court identified newspapers as an example of two-sided platforms in which this does not apply: ‘But in the newspaper-advertisement market, the indirect networks effects operate in only one direction; newspaper readers are largely indifferent to the amount of advertising that a newspaper contains.’ 

It is yet unclear exactly where any particular online, advertising-based business operating a multisided platform will fall on the Supreme Court’s spectrum – closer to Amex’s payment network or to a newspaper. That will be highly fact-specific. But it is worth evaluating that, in each particular case, including in the merger context, to ensure that the full impact of a merger and a complete set of competitors of the merging parties are taken into account when in an agency review. The director of the FTC’s Bureau of Competition, Bruce Hoffman, recently expressed significant interest in ‘litigating the parameters’ of the American Express decision, and that he is ‘very interested in bringing two-sided market cases’. The outcome of these hearings and future cases involving multisided platforms will show how significant a factor (or not) Ohio v. American Express will play in merger reviews involving digital markets.


While unlikely to become law, the Better Deal bill and its proponents may have been a contributing factor in agency officials committing resources to issues raised by the bill. The FTC’s launch of its Technology Task Force, Chairman Simon’s focus on acquisitions of nascent competitors by digital platforms, and the topics of the FTC’s Hearings certainly suggest so. The FTC Hearings involved more than 350 panellists and over 850 public comments, and covered topics such as ‘competition . . . issues in communication, information, and media technology networks, . . . markets featuring “platform businesses,” . . . intersection between privacy, big data, and competition . . . [e]valuating the competitive effects of corporate acquisitions and mergers’, and ‘monopsony power, including but not limited to, in labor markets’. 

That said, the agencies have already thoroughly investigated mergers involving digital markets and big data issues many times before. And merger enforcement has picked up quite a bit since the Obama administration, with several successfully litigated mergers, including in tech markets (e.g., Bazaarvoice/PowerReviews). Thus, unless merger retrospectives uncover compelling evidence of significant consumer harms from past under-enforcement, the question is how much more aggressive or experimental the administration will want to or realistically can be in its merger enforcement (other than perhaps being stricter on remedies). Doing so could risk more harm than good and run into boundaries set by judicial precedent (especially with the significant number of conservative judges that have recently joined the bench, including at the appellate court levels). The outcome in DOJ’s recent challenge of the AT&T/Time Warner merger is illustrative in that respect. The European Commission has greater discretion in that sense, since it does not have to prevail in court to stop a merger, and typically is given more deference by the courts than are the US agencies. Yet, despite its historically greater policy focus on big data and tech issues, even the European Commission has not to our knowledge blocked mergers based on concerns about such issues.

It looks as though the current administration, at least at the FTC, will devote more investigative resources to mergers in these areas of the new economy, especially acquisitions of start-up companies in the technology sector, but it remains to be seen whether that will lead to much more enforcement. That may depend, in part, on the outcome of the FTC Hearings and merger retrospectives, as well as political considerations.

The one area in which there has been significant change in merger enforcement policies compared to the last administration, is in remedies, especially in vertical mergers. That difference could surface in merger investigations involving digital markets and big data, since non-horizontal theories of harm are quite common there. As discussed, the DOJ’s loss in AT&T/Time Warner could mean it ends up intervening less in vertical mergers. On the other hand, the agencies clearly have been serious about not favouring behavioural remedies and typically insisting on divestiture remedies even in vertical mergers. That is something to take into account when planning vertical mergers that could raise significant opposition, including when negotiating the antitrust risk allocation in merger agreements. The agency’s policy position on behavioural remedies, combined with the outcome in the AT&T/Time Warner case, could in some circumstances cause parties to negotiate longer drop-dead dates, more litigation commitments and greater reverse break fees in merger agreements for such deals.

Subscribe here for related content, breaking news and market analysis from Global Competition Review.

Global Competition Review covers the most crucial developments in competition law and enforcement worldwide with breaking news commentary and analysis five days a week, as well as original in-depth reports, interviews and features on antitrust enforcement worldwide.