Overview: Competition Economics

Digital platforms on the rise

Digital platforms are increasingly important to the economic fabric of the Asia-Pacific as the use of the internet, social media and matching and sharing applications expands. Consumers across the region increasingly use digital platforms to stay in touch with each other, and to buy goods and services.

Reflecting a global trend, the competitive effects of digital platforms are the subject of intensifying scrutiny across the region. Several antitrust cases involving digital platforms have been brought in the region over the past 12 months, and competition authorities have frequently drawn attention to the potential for anticompetitive outcomes in digital platform services.1 The Economist has opened 2018 highlighting the rapid evolution of digital platforms as key for economies around the world.2

There are many indications that digital platforms will continue to be a growing focus for competition authorities in 2018. In this overview, we identify the essential economic characteristics of digital platforms and consider how the traditional frameworks and tools of competition analysis may need to be adapted in light of their evolving prominence.

Characteristics of digital platforms

Digital platforms are online marketplaces that allow multiple user types to access services for which there is direct interaction between at least two of the user types. For example:3

• some Facebook users access a social media network, while other users place advertisements that are seen by those using the social network;
• Uber matches vehicle drivers with passengers requiring a point-to-point transport service; and
• eBay operates a digital platform where users can buy and sell products and services.
  Some characteristics common to most digital platforms affect the analysis of competition issues, such as:

• the value of a digital platform to one type of user often increases with the numbers of another type of user – a phenomenon known as ‘indirect network effects’.4 For example, drivers will value Uber’s services more highly the more that would-be passengers use its platform. Digital platforms with a larger user base are often better able to provide their services than a platform offering the same service but with fewer users;
• these network effects, in combination with their economies of scale, mean that digital platforms can grow very quickly, much more so than many traditional ‘bricks and mortar’ providers of goods or services;
• the user of digital platforms often creates large amounts of valuable data, eg:5
  • Uber collects information on its users’ locations, their transactions for its ride sharing service and the devices they utilise;6
  • Facebook collects information on the content provided by its users, the connections between them, on financial transactions conducted on Facebook, the devices used to access its platform, and information from websites and apps that use its services;7 and
• consumers often use multiple digital platforms for the same service (known as ‘multi-homing’).
• However, digital platforms also exhibit many different characteristics. For example:
  
• some platforms facilitate transactions between users, while others provide services directly to users, for example:8
  • matching platforms such as Airbnb, Uber and eBay allow users to find and purchase services and products, often earning a commission on those transactions;
  • other platforms attract users for some intrinsic purpose, thereby providing the basis for advertising to take place (eg, Facebook, Twitter, Spotify and Google);
• some digital platform providers focus on one type of service, while others offer a wide range of services, potentially over a number of interconnected platforms;9 and
• platforms have a range of pricing structures and levels, with some services being offered for free.

The business models of some digital platforms have changed significantly over time. TripAdvisor initially provided a means for customers to review hotels and other services. It then expanded to offer metasearch, where TripAdvisor would aggregate information from a number of different websites in response to searches for accommodation, with a heavy reliance on advertising.10 Later, TripAdvisor introduced a feature allowing users to book hotels and other services, charging accommodation providers a commission/booking fee.

Market definition

Competition analysis must inevitably focus on what goods or services are being provided and the firms that compete to do so. Establishing the bounds of the relevant market provides a practical framework for identifying the products that consumers regard as close substitutes for those that are of interest, having regard to their characteristics, intended use and location.

Defining the market in which any digital platform service is offered begins in much the same way as with any other circumstance – by describing the products or services being provided. Digital platform services may be described in terms of their offering to one or other type of user, such as:

• the ability for would-be passengers to find a point-to-point transport service; and
• the ability for vehicle drivers to find passengers.

Alternatively, digital platforms may be described as intermediary services that are provided to multiple user types, eg, the service of matching passengers to drivers. Generally, the stronger the network effects, the more likely will be the case for considering potentially separate services as part of the same market, because the services are more closely intertwined.

The type of service offered may also be relevant to market definition, with platforms providing audience-like services (eg, Spotify) tending to have more limited network effects. Platforms themselves also benefit from indirect network effects, because their value to advertisers increases with the number of potential user ‘eyeballs’. However, users generally do not value the platform more highly when there are more advertisements.

Weaker network effects may mean that markets should be defined in terms of the separate services offered to different types of users, rather than at a platform level.

The principal determinant of the markets in which any digital platform operates remains the same as for any product or service, ie, the demand- and supply-side substitutability of the services it provides. The former can be assessed by reference to the extent customers can be expected to switch to other services if prices were to increase or quality were to reduce – noting that the price to some users may be zero.11 Network effects may mean that when some users switch away from a particular platform, the value of the services provided to other users may fall, causing yet more users to switch away. This effect must be taken into account in the assessment of demand-side substitution.12

One likely question is whether the services provided by a digital platform are in the same market as more traditional bases for the supply of a similar service. The answer is likely to vary from case to case. For example:

• in a recent decision, the Competition Commission of India (CCI) said that traditional forms of messaging such as mobile phone texting and calling are not in the same market as instant messaging services using smart phones, primarily because the latter forms of messaging have:13
  • different qualities, ie, instant messaging applications are restricted to use on smartphones, they must be connected to the internet and they provide additional functionalities to users; and
  • different prices, ie, instant messaging applications are generally free; and
• in contrast, the CCI recently rejected the contention made by Ola Cabs (a digital ride-hail platform offering services similar to that of Uber) that it was only a technology company and does not compete in the same market as those firms that own taxis14 – the CCI said that adopting a new business model to provide the same services did not put Ola Cabs in a separate market.15

The question of whether other suppliers can quickly offer the service in question without substantial investment – the extent of supply-side substitution – may also be relevant. In some instances, it may be easy for an operator to provide an additional service to a large number of existing customers, although these may or may not be successful. For example:

• Facebook introduced a mobile instant messaging service in 2014, and now has more than 1.2 billion monthly users;16 but
• Google introduced Google Plus in 2011 in an attempt to create a social network, but this largely failed.17

These contrasting examples underline the challenges in assessing, a priori, whether a particular supply-side expansion is likely to be successful.

Assessing market power

Assessing the extent to which a firm has substantial market power or – in European terms – is dominant is generally a prerequisite for the analysis of unilateral conduct issues, while the potential for any market power to be strengthened sits at the heart of merger analysis.

The competitive constraints that curb the potential market power of digital platforms are of essentially the same form as those faced by any other firm, ie, existing competitors, new entrants and the potential for one or more user types to switch to an alternative provider. But digital platforms also give rise to some important distinctions in the analysis of market power, as compared with more conventional forms of market.

Network effects may have two influences on the potential for market power; ie, they may:

• strengthen the market power of a firm with a relatively large market share, because it has the most users and so the best quality service (holding all else equal); but 
• encourage ‘step change’ improvements in the service of small but growing platforms, and vice versa.
• The extent to which network effects may affect market power depends upon:
• differences in the scale of competing digital platforms;
• the value that users place on there being a large number of other users on the network; and
• whether there are network effects on one or both sides of a digital platform.

Many platforms benefit from strong economies of scale, leading to lower costs for large firms, which can benefit consumers, but may also increase barriers to entry.

The market power of a platform will often depend on whether there is multi-homing, thereby allowing users to switch quickly between platforms and increasing competition between them.

There is a fast pace of innovation and disruption in digital platforms, and so high market shares may not reflect enduring market power because even a large firm can be quickly overtaken by new technology. However, competition authorities will not usually accept ‘[a]bstract vulnerabilities expected at some unspecified point in the future’,18 and it is intrinsically challenging to provide evidence of a technology or innovation that has yet to be developed.

Digital platforms often hold substantial quantities of data, which can be used to provide better services. But this same phenomenon may mean that smaller firms unable access the same (or similar) data will find it more difficult to compete with a large-scale incumbent. In assessing the extent to which data may give a firm market power, relevant considerations are likely to be the scale of the advantage that the data provides the incumbent, and the degree to which this is replicable by others (by using similar or different data).

Exclusionary conduct

Exclusionary conduct arises in several potential forms. Predatory pricing and exclusive dealing have been the most common of unilateral conduct matters involving digital platforms in the Asia-Pacific region in 2017.

Two predatory pricing matters were investigated in India, but neither was found to be anticompetitive. In one instance, the CCI accepted, in part, that the low prices were not because of cost efficiency,19 but rather facilitated a big expansion in a short time so that network effects would be realised.20 In other words, network effects were accepted as justification for below cost pricing in some circumstances. The CCI also said it was hesitant to intervene in a market that had not yet fully evolved, since this it may lead to sub-optimal outcomes.21

Some matters have alleged exclusive dealing by digital platforms, usually that customers are not allowed to trade with a rival platform.22 It seems likely that there will be multiple investigations pursued or continued in 2018, including in Japan (investigating Airbnb), South Korea (potentially investigating eBay) and China (potentially investigating Alibaba).23

Similarly, some digital platforms have been accused of foreclosing rivals by preferring their own websites or apps. The most notable of these is the European Commission’s June 2017 decision to fine Google €2.42 billion for giving illegal advantage to its own comparison shopping service.24

Similar allegations/investigations have been made in the Asia-Pacific region in 2017 and are likely to continue during the coming year. For example, Google has been accused of leveraging its dominance in the online search market in India,25 and Naver in South Korea26 was accused of favouring its own housing website links over those paid for by advertisers.

Anticompetitive conduct involving ‘big data’

For some time now, various concerns have been expressed to the effect that big data may be used anticompetitively. Although many competition authorities, including in the Asia-Pacific, apparently still hold some concerns, there are few examples in which the abuse of a dominant position has arisen from the use of data.

The principal concerns expressed in relation to the use of big data are that:27

• the ability of a platform to collect and use data, together with network effects, leads to that platform having more data and growing larger, thereby giving incumbents an advantage over smaller firms that cannot easily access similar data;28
• firms with big data may prevent smaller rivals from accessing their data, thereby increasing switching costs for users;29
• big data may facilitate price discrimination, through its deployment in setting high prices for customers with a high willingness to pay for a service, and vice versa;
• mergers between firms with data may increase their market power; and
• big data may be used by algorithms leading to a collusive outcome.30

One of the difficulties with these concerns is that acquiring and using big data also has pro-competitive effects, since it generally allows firms to offer better services. Further, data can often be acquired or accumulated by many different parties.

In more proactive initiatives involving big data, a number of countries have taken action that may reduce the potential market power arising from data. For example:

• the General Data Protection Regulation (GDPR) in the EU will come into force this year, and gives consumers the right to access their own data (along with other rights);31 and
• the Productivity Commission in Australia recently set out a range of recommendations in relation to data availability and use, including that consumers should have the right to access and use data about themselves, including to give it to a third party.32

These recommendations may well make it easier for some smaller digital platforms to access the data they need to provide high-quality services to their customers, and so may limit the intrinsic advantages of big data held by incumbent firms.

Mergers

The presence of network effects and economies of scale will generally heighten concerns arising in relation to mergers between rival digital platforms, again because these attributes may make it more difficult for smaller firms to exert a strong competitive constraint on a larger firm.

The unconditional approval of the Facebook/WhatsApp merger by the European Commission in 2014 was an important decision for digital platforms. The Commission found that traditional electronic communications such as texts and phone calls did not substantially constraint consumer communications apps, because those apps had a greater functionality and were generally free.

Reflecting many of the challenges described above, the Commission had a number of difficulties in estimating shares of the relevant market, including that:

• market shares by revenue did not make much sense when the products were free of charge;
• there were no obvious volume-based measures of market shares; and
• market shares fluctuated substantially over time.

The Commission also found that there were network effects in the market for consumer communications apps, but a number of factors mitigated their negative effect on competition. In particular:

• barriers to entry in terms of cost and time were low, and new apps were constantly entering the market;
• there was extensive multi-homing; and
• the users 
• remained in control of their phones and could quickly and easily install a new communications app, and use their existing network of contacts on WhatsApp.

Late last year the merger of two more traditional media organisations, NZME and Fairfax, was blocked in New Zealand.33 The Commerce Commission said that providers of digital advertising, such as Google and Facebook, constrain the merging parties,34 and that the proposed merger would not have the effect of substantially lessening competition in the market for digital advertising.35 However, the Commerce Commission also defined some print advertising markets in which digital platforms did not compete.

In relation to news content, the Commerce Commission said that digital platforms such as Facebook and Google were operating only at a downstream distribution level, and were not involved in the production of news content,36 whereas many traditional media companies operate both upstream and downstream.

The investigation of the proposed collaboration between Uber and CDG Taxis by the Competition Commission Singapore will be an interesting matter to monitor as this year unfolds.

Conclusion

The fundamental principles that competition authorities should apply (and generally are applying) to services provided by digital platforms are the same as those for any other industry. However, the unusual characteristics of digital platforms must be taken into account to determine how various forms of conduct may affect competition between such platforms. These characteristics differ between platforms and over time, so it will always be important to understand the particular facts of each matter, on a case-by-case basis.

There is likely to be a range of initiatives by competition authorities this year, in light of the challenges posed by digital platforms. At one end of the spectrum, the Hong Kong Competition Commission has said that it will not examine the major platforms such as Amazon and Google,37 while at the other end, the Australian Competition and Consumer Commission (ACCC) has just commenced a detailed inquiry into the effect that digital search engines, social media platforms and other digital content aggregation platforms have on competition in media and advertising services markets.38

The authors are grateful for the assistance of Brydon McLeod in preparing this article.

Notes

1 For example, see: Li, T, H, How Competition Authorities can stay relevant in the digital economy, Prepared for the Asean Competition Conference 2017, March 2017; Sugimoto, K, New Year Message from Chairman Sugimoto, Japan Fair Trade Commission, January 2017; and Sims, R, The ACCC’s approach to colluding robots, Can robots collude?, November 2017.

2 The Economist’s 20 January 2018 leader, ‘How to tame the tech titans’ noted some of the competition challenges presented by Google, Facebook, Amazon and other big digital platforms, and identified potential remedies.

3 A number of different definitions of platforms have been proposed. See, for example: Bundeskartellamt, B6-113/15, Working Paper – Market Power of Platforms and Networks, June 2016, pp 8-13.

4 The value to one type of user may also increase as another type of user interacts more intensely with the platform, eg, the same number of users access the platform, but for a longer period of time.

5 The Economist explained that data is a lucrative commodity, made abundant by smart phones and the internet in a leader article from 2017. See: The Economist, 'Regulating the Internets Giants: The World’s most valuable resource is no longer oil, but data', May 2017.

6 See: https://privacy.uber.com/policy, accessed 25 January 2018.

7 See: https://www.facebook.com/policy.php, accessed 25 January 2018.

8 Bundeskartellamt, B6-113/15, Working Paper – Market Power of Platforms and Networks, June 2016, p 18.

9 Some digital platforms have been expanding their products and services. For example, Google (into mobile phones), Amazon (into Ebooks), Uber (into food delivery) and LinkedIn (into recruiting).

10 See TripAdvisor, 2014 Annual Report and notice of 2015 Annual Meeting and Proxy Statement, April 2015, pp 4-6.

11 Markets for services provided over digital platforms have been defined, even if the service is provided for free. For example, see: European Commission, Case No. COMP/M.7217, October 2014, para 4.1.2; and Competition Commission of India, Case No. 46 of 2014, September 2014, para 16.

12 The standard version of the small but significant and non-transitory increase in price (SSNIP) test cannot be applied in relation to digital platforms because it does not take into account changes in volume on more than one side of a market. Modified SSNIP tests have been proposed, eg Filistrucchi, L, A SSNIP Test for Two-Sided Markets: the case of the media, NET institute working paper No. 08-34, October 2008.

13 CC, Case No. 99 of 2016, June 2017, para 11.

14 CC, Case No. 6 & 74 of 2015, June 2017, para 9.

15 CC, Case No. 6 & 74 of 2015, June 2017, para 62.

16 Forbes, Facebook Messenger Passes 1.2 Billion Users, April 2017.

17 See Baldoni, J, Why Google Plus is a Failure, CBS.

18 Bundeskartellamt, B6-113/15, Working Paper – Market Power of Platforms and Networks, June 2016, p 74.

19 CC, Case No. 6 & 74 of 2015, July 2017, para 122.

20 CC, Case No. 6 & 74 of 2015, July 2017, para 122. The CCI also concluded that the firm setting the low prices was not dominant.

21 CC, Case No. 6 & 74 of 2015, July 2017, para 123.

22 Meituan, an online food-delivery service in China was found guilty of having abused its dominance because it had asked restaurants to commit to using only its services (by signing a cooperation commitment letter). See PaRR, China’s Jinhua regulator fines Meituan for restricting competition (translated), August 2017.

23 See: Asia Nikkei, Airbnb probed for antitrust violations in Japan, November 2017; The Hankyoreh, Ministry of SMEs and Startups recommends investigation of eBay Korea, November 2017; and PaRR, JD.com files antitrust suit against Alibaba in Beijing court sources, January 2018.

24 European Commission, Antitrust: Commission fines Google €2.42 billion for abusing dominance as search engine by giving illegal advantage to own shopping service, June 2017.

25 The allegations include the denial of access and imposition of unfair conditions on ‘AdWords’ customers, and that it favoured its own products in search results. See: Indian Express, Google abused dominance in India, finds Competition watchdog probe, September 2015; and Patel, F, India on course for Google decision within nine months, hearings over, PaRR, January 2017.

26 See PaRR, Korean antitrust authority to look into Naver (translated), October 2017.

27 There are a range of potential concerns regarding privacy too that we do not discuss here.

28 See Lambrecht, A, and Tucker, C, 'Working Paper - Can Big Data Protect a Firm from Competition?', December 2015.

29 See Stucke, M, and Grunes, A, 'Big Data and Competition Policy', Oxford University Press, England, June 2016.

30 See Stucke, M, and Ezrachi, A, Virtual Competition: The Promise and Perils of the Algorithm-Driven Economy, Harvard University Press, Cambridge, November 2016.

31 See GDPR Key Changes, available at: https://www.eugdpr.org/key-changes.html, accessed 25 January 2018.

32 Productivity Commission, Data Availability and Use: Inquiry Report, March 2017, p 35.

33 The Commerce Commission declined to grant authorisation for the merger, and subsequently, the High Court dismissed an appeal of that decision. See New Zealand Commerce Commission, NZCC 8 - NZME Limited and Fairfax New Zealand Limited – Authorisation Determination, May 2017.

34 Ibid, para 317.

35 Ibid, para 325.

36 Ibid, para 575.

37 See: PaRR, Hong Kong lays off Google Amazon, citing agency inexperience – LexMundi- Deacons conference, June 2017.

38 ACCC, Digital platforms inquiry: Project overview, available at: https://www.accc.gov.au/about-us/inquiries/digital-platforms-inquiry.