The role of data and how it can harm competition is a recurring topic in contemporary antitrust literature. Below, we highlight a few points integral to understanding what the fuss is all about.
Making sense of data
Frankly speaking, data itself is not revolutionary, and it presents nothing that competition authorities have not seen before. Indeed, datasets such as customer lists, purchase histories and other data describing consumer behaviour have been collected and utilized since the beginning of business itself. It is only the ways in which data can be collected and analyzed that has changed along with the developments in automation and data science over the past years, making exponential volumes of data available to an increasingly wide range of industries that are able to use said data in ways unprecedented before.
While this type of data might be a stranger in the field of competition law, it is an old friend of intellectual property experts. Data in the sense of proprietary information are all but universally protected as trade secrets. In certain jurisdictions data may be given copyright protection, and the methods and tools for access and analysis are often guarded by patents. Within modern intellectual property regimes, data can prove invaluable. As a trade secret – much like the recipe behind Coca Cola – data can be held out of the hands of competitors in perpetuity and use based on even serendipitous discovery can be sanctioned. Whereas copyrights and patents are time-bound, protection periods are longer than the meaningful lifespan of the underlying data, allowing proprietors to throttle the use of data as they please.
As proven by numerous pharmaceutical disputes, the crossroads between intellectual property rights and competition is fertile ground for discontent. In spite of their professed differences, the disciplines share much in common. Most importantly, both competition and intellectual property policies aim at generating social gains through dynamically efficient and innovative markets. Against this backdrop, both disciplines hold data as an input that is valuable to those who hold it, but whose use must also benefit the society at large.
The true value of data stems from its dynamic effects on proprietors and society. Consequently, the crux of concerns caused by the collection and processing of enormous volumes of data in nearly endless ways, is in the risk of, inter alia, foreclosing competition – not in the use of data itself. In assessing the risk of foreclosure, we must also consider the existence of market power.
Does data create market power
In the past year national competition authorities have provided some rough guidance to help determine when the collection and possession of data may generate market power. The main requirements are that:
- data has a significant role as an input in making goods or performing services (scale or scope advantagesA certain amount of data may allow more refined and better-targeted services. Collection of a certain amount of data may also enable expansion into adjacent markets.);Graef, Inge, ‘Market Definition and Market Power in Data: The Case of Online Platforms (September 8, 2015)’, forthcoming in World Competition: Law and Economics Review, Vol. 38, No. 4 … Continue reading and
- competitors cannot reasonably collect, replicate, purchase or substitute such data.See e.g. CMA (2015): The Commercial Use of Consumer Data, p. 89 & Autorité de la Concurrence & Bundeskartellamt (2016): Competition and Data, p. 47-54. See also relevant case law: Cisco … Continue reading
Hence, as long as the tools and markets support rivalry, the fact that rolodexes and dusty binders have been superseded by voluminous databases and advanced algorithms should not fundamentally change the way in which data is viewed.
Even if the above criteria are fulfilled, data is rarely an essential facility – a question occasionally entertained in legal literature. It is worth remembering that the threshold for treating an input as essential is remarkably highSee ECJ in IMS Health (case C-418/01), paras 34-52 & Bronner (case C-7/97), paras 44-45. and the remedies involved, including sharing data and compulsory licensing, have their own additional problems concerning privacy, incentives to innovate and more.
However, in absence of fulfilling the above criteria, it is also worth noting that data may also have the effect of diluting market power. In traditional and more static industries, market shares are a more reliable indicator of market power and entry may require significant monetary investments regardless of any efficiency gains that data may bring. In digital markets however, market shares may not be as reliable. In fact, if data is a substantial input and the particular data used is replicable,Instances of replicable data include user demographics and preferences and current-location related data. any competitive advantage may be undermined by competitors who gather and use data differently, or provide different services based on the same input. In such cases, data may help lower entry boundaries and actually allow more dynamic competition.
Making merger notifications work in a data-driven world
Current EU rules on merger notifications are inadequate when a transaction revolves around data. Transactions may involve target businesses whose turnover or market share do not reflect their true competitive significance. However, since European notification regimes are triggered based on turnover or market share, significant mergers may currently be avoiding merger notification and ex ante review by competition authorities entirely.E.g. non-traded data often is particularly important on the subsidized side of a two-sided market. Lamadrid (2014): On Privacy, Big Data and Competition Law (Post 1/2), available at: … Continue reading A material, if not recent, example is the Facebook/WhatsAppFacebook/WhatsApp (case COMP/M.7217). merger in which the concentration did not meet the turnover thresholds to be notified to the European Commission, despite its clear pan-European significance.The Commission did eventually get to decide on the case, owing to referrals from Spain and Portugal who maintain market share based thresholds for merger notifications.
The leading resolution seems to be introducing thresholds based on transaction value. Proposed by the German Monopolies Commission in 2014Monopolies Commission (2015): Competition Policy: The Challenge of Digital Markets, para S54. and recently echoed by EU Commissioner Vestager by stating that “…the value of a merger could be a good guide to its importance”,Commissioner Vestager (2016): Refining the EU Merger Control System, available at: https://ec.europa.eu/commission/2014-2019/vestager/announcements/refining-eu-merger-control-system_en. it seems as this approach could help merger control function in the realm of data-driven markets. The frontrunner is Germany, in which a legislative change is already being planned to include a threshold of MEUR 350 triggering a notification obligation. Germany’s plan is that the change is effective before year’s end.The German Federal Ministry for Economic Affairs and Energy (2016): Green Paper on Digital Platforms, p. 49. It will be interesting to see whether the rest of Europe follows suit.
By and large, the collection and use of data presents great possibilities for innovation and ultimate benefits to consumers. It can also be a source of significant market power, which may result in abusive conduct, unless a clear and predictable framework exists.
However, we should not be overthrowing competition policies just yet.One should tread with care when considering sector specific regulation within an established field of law. See Eastbrook (1996): Cyberspace and the Law of the Horse, available at … Continue reading Rather, we should be interpreting them in ways that include to the characteristics of fast-paced data markets where data has numerous different forms and uses. Many of these uses are not only beneficial for companies themselves, but they are also often pro-competitive.
Accordingly, competition authorities must guide firms consistently and provide a clear framework that concurrently fosters dynamic competition. Innovation in data-driven markets has remarkable potential and should not be unnecessarily hindered.
The views expressed in this post purely reflect those of the authors alone.
|↑1||A certain amount of data may allow more refined and better-targeted services. Collection of a certain amount of data may also enable expansion into adjacent markets.|
|↑2||Graef, Inge, ‘Market Definition and Market Power in Data: The Case of Online Platforms (September 8, 2015)’, forthcoming in World Competition: Law and Economics Review, Vol. 38, No. 4 (2015), available at SSRN: http://ssrn.com/abstract=2657732.|
|↑3||See e.g. CMA (2015): The Commercial Use of Consumer Data, p. 89 & Autorité de la Concurrence & Bundeskartellamt (2016): Competition and Data, p. 47-54. See also relevant case law: Cisco Systems and Messagenet v European Commission (case T‑79/12), paras 65-84; Facebook/WhatsApp (case COMP/M.7217), paras 184-189.|
|↑4||See ECJ in IMS Health (case C-418/01), paras 34-52 & Bronner (case C-7/97), paras 44-45.|
|↑5||Instances of replicable data include user demographics and preferences and current-location related data.|
|↑6||E.g. non-traded data often is particularly important on the subsidized side of a two-sided market. Lamadrid (2014): On Privacy, Big Data and Competition Law (Post 1/2), available at: https://chillingcompetition.com/2014/06/03/.|
|↑7||Facebook/WhatsApp (case COMP/M.7217).|
|↑8||The Commission did eventually get to decide on the case, owing to referrals from Spain and Portugal who maintain market share based thresholds for merger notifications.|
|↑9||Monopolies Commission (2015): Competition Policy: The Challenge of Digital Markets, para S54.|
|↑10||Commissioner Vestager (2016): Refining the EU Merger Control System, available at: https://ec.europa.eu/commission/2014-2019/vestager/announcements/refining-eu-merger-control-system_en.|
|↑11||The German Federal Ministry for Economic Affairs and Energy (2016): Green Paper on Digital Platforms, p. 49.|
|↑12||One should tread with care when considering sector specific regulation within an established field of law. See Eastbrook (1996): Cyberspace and the Law of the Horse, available at http://chicagounbound.uchicago.edu/cgi/viewcontent.cgi?article=2147&context=journal_articles.|