It is expected that knowledge derived from big data analysis will inspire further innovation across existing industries in the context of the spread of the internet of things and the advancement of artificial intelligence (AI) technology. As data becomes more and more important in business activities, the establishment of an environment of free and fair competition to collect and utilise data is essential to spur innovation. Competition authorities may face challenges that they have never before confronted.
We, the Japan Fair Trade Commission (JFTC), released the remarks by Chairman Kazuyuki Sugimoto in July 2017 for the celebration of the 70th anniversary of the Japanese Antimonopoly Act (the Act), to promote the significance of ensuring free and fair competition in the market. In the remarks, he mentioned that we have been working vigorously to clarify competition policy on data and IT. He emphasised the impact on competition of the accumulation and utilisation of data and the enhancement of information-gathering capabilities in fields related to IT and digitisation. For example, we published a report on data and competition policy in June 2017. The report contained the discussion held in the Study Group on Data and Competition Policy.The Study Group was organised by the Competition Policy Research Centre (CPRC) of the JFTC.
The report drew a distinction between personal data and industrial data. Personal data referred to information related to individuals and the industrial data referred to information collected with certain sensors concerning the status of actual tangible things such as devices.
While incumbent digital platform service providers can collect and accumulate a mass of personal data by offering ‘free’ services or networks, new entrants cannot do so because of the cost of constructing a network to collect the personal data.
On the other hand, the significance of industrial data is highly dependent on circumstances. Industrial data is valuable only to those with knowledge and expertise regarding the operation of the relevant devices. At the same time, they often want to retain control over that industrial data and are not likely to share it. For those reasons, an unnecessary level of data hoarding is likely to happen, reflecting the will of the parties concerned. Accordingly, utilisation of industrial data has not been much developed.
Basic approach to collection and utilisation of data
Today, the scope of the use of data is expanding and value of data is soaring. While bearing in mind that this has promoted efficiency and innovation, we need to promote competition among the enterprises that are engaged in collecting, accumulating and utilising data.
We should therefore eliminate anticompetitive conduct that may impede competitive business activities, and block mergers that may generate a concentration of the data in a small number of enterprises, so that the utilisation of data in an unreasonable manner does not restrict competition in a data-related market.
The report also discussed other issues, for example, the portability of personal data, ownership of industrial data and open access to public data.
Methods of analysing the effect in reducing competition
The Act prohibits certain conduct that reduces (or might reduce) competition. It requires a comprehensive assessment to determine whether competition in a market can be reduced by the accumulation and utilisation of data. That may include many aspects such as:
- the nature and form of the alleged conduct;
- the existence of a competitive relationship between parties with respect to that conduct;
- the position that each party occupies in the market;
- the conditions in the market as a whole;
- whether there are reasonable grounds for restrictive practices; and
- the effect on the willingness actively to accumulate and utilise data.
We should be careful that the accumulation of data by a specific enterprise does not make it difficult for other enterprises to enter or remain in the market, which could contribute to the establishment, maintenance and strengthening of market power. That is especially the case when there is a huge difference in the ability of competitors to collect raw data.
Basic approach to relevant markets
We should also examine the impact of anticompetitive conduct on competition in the market when we assess whether it reduces competition. Markets are defined from the standpoint of substitutability for buyers and, if necessary, for suppliers.
The way to define a data-related market is not so different from other markets. However, we should pay attention, when we analyse the effect of the conduct, to certain factors that specifically characterise data-related markets such as:
- that there is little limit on transportation of data;
- that data has a close relationship with R&D;
- that there are multi-sided markets in the data transaction; and
- that services are often free-of-charge to users.
The anticompetitive conduct raised in the report
The report classified potential anticompetitive conduct into two types: those related to data collection and those related to access to the collected data.
Conduct relating to the collection of data
Collection by a single enterprise
Enterprises sometimes conclude confidentiality agreements when they enter into joint R&D projects, under which one party may acquire all the rights concerning the data and technology obtained in the process of the joint R&D. That requirement may have the effect of strengthening the position of that party in the market, particularly if the data is rare. It may impede other parties’ R&D activities and ultimately lessen competition in the market. Accordingly, the conduct may violate Article 19of the Act, which prohibits unfair trade practices, if it impedes fair competition or if it constitutes an abuse of a superior bargaining position. The latter may arise if the party is in a superior position in the transaction with others, and the former gives unreasonable disadvantage to the latter.
In addition, platform markets sometimes have ‘lock-in effects’. When a platform service provider has certain power in its market, it may be difficult for customers to switch the provider to others even if the provider changes terms and conditions in a way that disadvantages customers. In such a case, the conduct may be in breach of Article 3 (Private monopolisation) or Article 19 of the Act.
Joint collection of data
Enterprises may collect data in cooperation with others when they consider that joint collection may improve efficiency and create innovation. That conduct is, in general, considered to contribute to competition.
However, it can fall under Article 3 of the Actbecause it may facilitate collusive conduct, as participants may acquire, in the course of the joint data collection, detailed information that is directly linked to the core of their business, such as cost, price, quantity and quality of the service.
Moreover, joint data collection also raises anticompetitive concerns if it restrains participants from collecting data that is essential for their activities.
The JFTC will take such factors below into consideration in determining the illegality of this conduct:
- the number of participants, market shares, etc;
- the nature of the data collected (for example, whether it is important for the R&D or as inputs);
- the necessity of joint collection; and
- the scope and period of the joint collection.
Conduct relating to access to the collected data
Recent technical innovation has made it possible for businesses to collect real-time data from various devices and use it to improve the quality of their services.
However, in some cases the use of real-time data on the operation of certain devices is limited. If the data is owned by an enterprise with market power in a device maintenance and inspection service market, and it does not permit its competitors to access the data, it will probably maintain and strengthen its control over the service market. At the very least, if the enterprise does not disclose the data to device owners when they request it, they may confront difficulties in doing their business. This conduct is called ‘data hoarding’.
In this context, any enterprise has the freedom to choose its trading partners and to set the terms of the transaction in their contracts, for instance, the period of time of disclosure. That does not breach the Act. The same applies to the accumulation of data. It does not normally result in a problem under the Act.
However, if some enterprises are not permitted to participate in the joint collection of data, or are allowed to participate but are unable to access the accumulated data, and find it difficult to engage in business activities as a result, that could infringe the Act.
Refusal of access by a single enterprise
Depending on the type of conduct, refusal of access to data may constitute exclusionary private monopolisationor refusal to trade. Moreover, if an enterprise obstructs its competitors from accumulating data, that can constituted interference with transactions.
Indeed, the refusal of access to the data without reasonable grounds can be deemed exclusive conduct. If it is executed ‘in a manner beyond the level of the normal competitive practices’ or if it amounts to a ‘refusal to trade as a means of achieving unreasonable objectives’, by an enterprise with market power, it is likely to infringe the Act, especially if the data is essential in the business activities of competitors, and it is technically or economically difficult for them to obtain alternative data. The conduct falls under the Act if it substantially restricts competition in the market or impedes fair competition.
Joint refusal of access
With regard to data gathered jointly by a number of enterprises with a relatively high total market share, if they prevent a specific enterprise from participating in the joint collection and do not allow the enterprise to access the date on reasonable terms, it may become difficult for the third party to carry out business activities, if they have no other means of obtaining the data. If there is a risk of them being excluded from the market in such a situation, those elements would probably constitute a violation of the Act in exceptional circumstances.
Even if participation in the joint collection of data is restricted, it needs to be considered that it may not become difficult for the third party concerned to carry out business activity, because the enterprise whose participation has been prevented could collect similar data independently.
Enterprises in a competitive relationship can collect data individually and pool the mutually substitutable data, and then collectively license the data to a third party (or refuse to do so). In that regard, the collective licensing of data through the data pool has aspects of both avoidance of competition among the enterprises participating in the data pool as to licensing (namely competition concerning data usage fees for data access) and the provision of multiple goods in combination, and it needs to be considered whether it reduces competition. There will normally not be an infringement of the Act if the importance of the data is slight, in the light of factors such as the degree of proliferation of products utilising the data; the existence of alternative data pools; the potential to license the data separately; and the potential for selective licensing. In general, it may constitute an infringement of the Act,if enterprises in a competitive relationship in a data-trading market jointly refuse to license data in the data pool and refuse to license data owned by each enterprise to third parties, without justifiable grounds, for the purpose of obstructing entry to a market of products utilising data as an input or excluding existing enterprises from the market.
Other conduct relating to data access that could be problematic under the Act include selling data bundled with other services such as analytical tools; obliging customers only to trade data with the enterprise concerned; or restricting the collection or use of data by parties other than the provider (including the owners of devices relating to industrial data) as a condition for providing technology such as machine learning technology (whether for free or for a charge). That conduct enables data to be used unreasonably and could be regarded as falling within a wide range of anticompetitive conduct, such as trading on restrictive terms and trading on exclusive terms.
Factors to be considered in reviewing data-related mergers
Some pointed out in the discussion that there are a number of conglomerate-type merger cases in which IT companies with mass data obtained through channels such as their own services enter completely different sectors, such as auto-driven car manufacturing and financial services, both domestically and overseas. Others observed that the number of data-related mergers had been increasing in recent years and that some digital platforms, already with large quantities of data, were attempting to buy up start-ups that could otherwise become their competitors in the future.
Given this situation, JFTC should pay attention to the following factors when it receives notification of a merger transaction in which merging parties have a large volume of accumulated data or have channels for collecting such data:
- privacy policies and the level of data protection;
- market power potentially affected by ongoing R&D;
- the impact of accumulated raw data by network and the analysis of data through algorithms; and
- the foreclosure effect on data supply.
Concerning pre-merger notification, notificiation thresholds based on turnover may not properly reflect merger cases in which it is difficult to foresee the impact of data on market power. A review of the thresholds may be necessary.
The Study Group concluded that the current framework under the Act is applicable to most competition concerns about the collection and utilisation of data. However, there remain issues not yet addressed such as digital cartels. The JFTC will pay attention to those issues and may have to deepen the discussion.
The JFTC also needs to be cautious about the monopolisation or oligopolisation of digital platform markets. The risk coincides with the convenience of those platforms.
The barrier of entry to the data-related market may be higher due to network effects as well as economies of scale and scope via the cycle of mass data collection, the improvement of quality of goods by AI, the concentration of customers and the leveraged expansion of network effects. Its impact may extend to other markets. We need to continue to consider the overall framework of competition regulation to facilitate entry to data-related markets.
 This chapter was prepared by the Secretariats of the Competition Policy Research Centre (CPRC) of the Japan Fair Trade Commission. The CPRC was established to build and improve functional and sustainable cooperative platforms between external researchers and practitioners, and the Japan Fair Trade Commission staff. The CPRC has reinforced the theoretical foundation on which the JFTC enforces Japanese competition law, and has made various suggestions and proposals for planning and evaluating competition policy from medium-and long-term perspectives, as well as using the platform to discuss current issues. The homepage of the CPRC is linked to JFTC’s website (https://www.jftc.go.jp/en/cprc/index.html).
 The report is available through JFTC’s website (https://www.jftc.go.jp/en/pressreleases/yearly-2017/June/170606_files/170606-4.pdf).
 Transaction with restrictive terms (Designation of Unfair Trade Practices, paragraph 2), https://www.jftc.go.jp/en/legislation_gls/unfairtradepractices.html.
 Unreasonable restraint of trade.
 Article 3 of the Antimonopoly Act.
 Article 19 of the Antimonopoly Act (Designation of Unfair Trade Practices, paragraph 2).
 Article 19 of the Antimonopoly Act (Designation of Unfair Trade Practices, paragraph 14).
 Article 3 of the Antimonopoly Act (private monopoly) and Article 19 of the Antimonopoly Act (Designation of Unfair Trade Practices, paragraph 2: other refusal of trade).
 Private monopoly, unfair trade practices (Article 2(9)(i) of the Antimonopoly Act).