Maziar Homayounnejad, PhD Researcher, King’s College London
Introduction
Recent years have seen a burgeoning of economic literature on multi-sided platforms (MSPs), regarding their peculiar features and their business models. Indeed since Rochet and Tirole’s seminal paper in 2003[1], there are now well over 200 peer-reviewed articles on the economics of these platforms[2]. Furthermore, as MSP business models often involve unorthodox and perhaps ambiguous pricing strategies, it is not surprising that competition lawyers have also taken an interest. Yet on both sides of the Atlantic there remains a dearth of legal analysis and policy in this area[3]. A recurring theme amongst lawyers is whether the issue of multi-sidedness is exclusively an evidential matter, i.e. for economists to deal with; or whether it also includes a jurisprudential element, whereby lawyers, competition authorities and the courts will have to adapt antitrust doctrines and perhaps create whole new ones to accommodate the growing significance of MSPs.
The following article will examine some of the main features of MSP businesses and suggest how these may be relevant for antitrust analysis.
The Economics of Multi-Sided Platforms
A MSP is essentially a business that serves “two [or more] distinct groups of customers who need each other in some way and who rely on the platform to intermediate transactions between them”[4]. Examples include video game consoles (where the two user groups are game players and game developers), dating clubs (men and women) and payment cards (consumers/cardholders and merchants) to name just a few. Table 1, below, contains an illustrative list of platform-based businesses and their various customer groups.
Table 1: Examples of platform-based businesses and their customers groups .
| Platform | Side A | Side B |
| Video game console | Game players | Game developers |
| Dating club | Men | Women |
| Payment card system | Merchants | Consumers/cardholders |
| Newspaper/TV station | Advertisers | Readers/viewers |
| Shopping mall | Shops/retailers | Shoppers |
| Estate agent | Vendors | House buyers |
| Financial exchange | Sellers | Buyers |
| PC operating system | PC users | Software developers |
| Traditional/online auction | Sellers | Buyers |
| Smartphone | ‘App’ developers | Smartphone users |
| Online search engine | Advertisers | Search users |
Fundamentally, these platforms facilitate value-creating exchanges in the face of transaction costs or other obstacles to trade, which would normally prevent the two sides from interacting[5]. According to Evans and Schmalensee (2007), they do this through the three core functions of MSPs[6]:
- Matchmaking – providing a physical location (e.g. shopping mall) or a virtual location (e.g. a website or other forms of electronic connectivity, like standardised payment cards and matching terminals) that brings buyers and sellers together;
- Audience-building – to provide liquidity and to make it more likely that an appropriate match will be found (see ‘The Importance of Critical Mass’ below);
- Shared resource to reduce common costs – as many of these platforms, like shopping malls and online auctions, involve very high set-up costs but negligible marginal costs of production. So, for example, rather than each retailer having his own security and customer parking, Westfield takes care of this and simply allows each retailer to hop onto its platform for an affordable fee/rent.
Indirect Network Effects
Perhaps more important for the application of competition law is the issue of indirect network benefits or externalities. In general, an externality occurs when economic agents “do not bear the full cost (benefit) from the harm (good) their actions do to others”[7]. Accordingly, some of the costs or benefits from an economic transaction affect third parties: these may be negative externalities, like pollution suffered by local residents near a factory; or positive externalities, like local residents benefiting from a rise in house prices when a major new employer sets up nearby. In both cases, the local residents suffering a loss or benefiting a gain do so as a spillover to the main economic activity, in which they were not necessarily involved.
Applying this concept to MSPs, we see, as alluded to above, that “the value that a customer on one side realises from the platform increases with the number of customers on the other side”[8]; accordingly, the demands of both sides are interlinked. For example, a search engine becomes more valuable to advertisers as there are more people using it to search the internet, because each one is a potential buyer. Likewise it is more valuable to search users if there are more advertisers on the platform, as this makes it more likely that they will see relevant adverts. Yet search users do not take into account that by using the platform they increase its value to advertisers, hence the benefits are external to them.
Importantly, the indirect network externalities are not solely determined by the number of customers on each side of the platform; in some industries at least, the quality of users on one side will also have a bearing on the level of benefit gained by users on the other side[9]. In this context, Hagiu[10] considers that ‘quality’ consists of any measurable characteristic or attribute held by one customer group, which increases the utility derived by the other. Such attributes are often objective, as seen in the example of LinkedIn, which only approves third party applications developers that are most relevant to professional social networking. The firm often excludes otherwise ‘high quality’ applications not fitting this profile[11]. Accordingly, Boudreau and Hagiu[12] consider that the firm may play a ‘public authority’ role by regulating access to the platform and, where necessary, to exclude ‘low quality’ users on one side of the market. Of course, the extent to which this happens depends on customers’ relative preferences for quantity as opposed to quality. What is fairly clear from an antitrust perspective, however, is that the significance of ‘quality’ for harnessing the indirect network effects in some MSPs may provide an objective justification in refusal to supply cases.
The ‘Chicken & Egg’ Problem and the Pricing Solution
Given that the platform must cater to at least two customer groups at the same time, MSPs have to work out how to get both parties on board without each one necessarily having immediate access to the other. This can be quite tricky and may require the platform to provide financial inducements to one party. A related issue concerns the need to balance the demands of both sides. Indeed the very purpose of the platform is to harness and ‘internalise’ the indirect network externalities, whenever customer groups are unable to do this on their own.
Accordingly, just as the government subsidises public transport and medicines on the NHS[13], the platform will also offer subsidised prices to one customer group (which may take the price below cost), while financing the subsidy through higher (perhaps significantly above cost) prices to the other group. The intuition, according to Evans and Schmalensee[14], is as follows:
- In a platform consisting of two customer groups, A and B, a fall in the price to members of B will get more of them on board. In a one-sided analysis, the extent to which this happens would depend solely on the price elasticity of demand for the product with respect to B customers only.
- However, since members of group A value the platform more if there is a greater number of Bs on there, more As will join the platform at the current price charged to A. The extent to which this happens depends on the indirect network benefit, which is measured by the value that As place on B’s membership; or, more formally, by the cross-price elasticity of demand for A goods with respect to the price charged to B customers.
- But with more As on the platform, Bs also value the platform more highly, leading to another rise in the number of Bs getting on board and so on.
- This ‘positive feedback effect’ may continue growing the business and giving the platform more critical mass .
Indeed even if Bs are not directly very price-sensitive, so long as the externalities between the two groups are strong, Evans and Schmalensee argue that participation by members of B can be highly sensitive to the price charged to them, and similarly for members of A. Thus as Rochet and Tirole[15] point out, there is a distinction between a price level and a price structure. Broadly speaking, the price level is the sum of the two prices, while the price structure is the ratio between them; and such is the importance of this to the multi-sidedness of a platform that Rochet and Tirole go as far as saying:
“A market is [multi]-sided if the platform can affect the volume of transactions by charging more to one side of the market and reducing the price paid by the other side by an equal amount; in other words, the price structure matters, and platforms must design it so as to bring both sides on board.”[16]
Accordingly, for a platform to be multi-sided it is both necessary and sufficient for the price structure to be non-neutral. More specifically, as alluded to above, MSPs are almost always expected to engage in a skewed pricing structure whereby one customer group may be charged below marginal cost, while the other is charged significantly above marginal cost[17]. Not only might this help to solve the chicken and egg problem, but it also helps to balance the demands of the two sides to ensure there is a sufficiently thick market as to make the platform viable[18]. This it does by offering the lower price to the customer group that is needed more, harder to acquire and/or more price sensitive; while the customer group that gets the most value from the other side of the platform and/or is less price sensitive often gets the higher price.
The profit-maximising formula for a two-sided market is, therefore, more complicated than for a one-sided business. While the latter will merely try to equate its marginal revenue with its marginal cost, the platform-based firm must consider a) the demands of both sides, b) the interrelationship between these demands (i.e. the network externalities), c) the costs directly attributable to each side (which may change as the volume of transactions alters) and d) the cost of running the platform[19]. Thus a fall in price to one side of the market, even if it goes below cost, is not necessarily predatory and may in fact be profit-maximising overall.
The Importance of Critical Mass
The existence of indirect network effects and the chicken and egg problem illustrates the importance of critical mass for platform-based businesses, particularly start-ups. Namely, the firm not only has to get members of both customer groups to appear on the platform, but also has to secure their participation in sufficient numbers to ensure there is an adequately thick and ‘liquid’ market. This is essential so that each customer group can maximise its chances of finding a value-creating match with members of the other side; only once this happens can customers justify the cost of participating in the platform and to subsequently return to it[20].
One further point can fairly be deduced from Evans[21]: that to keep improving the quality of matches and to remain a viable platform in the face of competition, balanced growth must be achieved fairly quickly. Otherwise, the platform will lose out to rivals who may grow faster and offer both user groups better quality matches in a more liquid market. The implication certainly seems to be that ‘bigger is better’ in the world of MSPs. Thus Evans points out: “just as more customers can propel a platform to success through indirect network effects, fewer customers can send a rival platform to its doom”[22]. This raises the danger of ‘tipping’ – that is, where the indirect network externalities in a ‘winner-takes-all’ market causes the whole market to tend towards a single dominant platform. The likelihood is perhaps aggravated in some high-technology markets like online search, where the cost structure – namely, very high fixed and sunk cost[23] but negligible marginal cost[24] – leads to economies of scale over a very large range of output. Indeed Shapiro and Varian[25] argue that when the network externalities and economies of scale in production combine, there is a ‘positive feedback’ effect, which sees strong firms getting stronger and more dominant, while weak firms get weaker. Of course this is less likely in markets where one or both user groups tends to multi-home, that is, to use more than one platform at the same time[26]. In such instances, the market will clearly have to accommodate several platforms and will face more competition.
A more immediate concern for the application of the competition rules is that such platforms may therefore engage in strategies to help them achieve liquidity while denying their rivals the chance to obtain the same critical mass[27]. The challenge for the competition authorities and the courts is therefore to distinguish between legitimate efforts by a platform-based business to harness the indirect network effects from an anti-competitive attempt to deny critical mass to its rivals. Namely, the familiar concept of ‘competition on the merits’ comes to mind here. In some cases, this may be assessed through a direct analogy with the rules developed for one-sided businesses, while in other cases there may arguably be a need to adapt those rules for MSPs.
[1]Rochet, J C. and Tirole, J (2003) ‘Platform Competition in Two-Sided Markets’, Journal of the European Economic Association, Vol. 1, No. 4, 990-1029.
[2] For a very thorough survey of the economics literature, see Evans, D S. and Schmalensee, R (2013) ‘The Antitrust Analysis of Multi-Sided Platform Businesses’. Available at SSRN: http://ssrn.com/abstract=2185373.
[3] Ware, R (2013) ‘Recent Developments in Two-Sided Markets in the USA and Canada’ Webinar, ABA Section of Antitrust Law.
[4] Evans, D S. (2008) ‘Competition and Regulatory Policy for Multi-Sided Platforms with Applications to the Web Economy’, pages 3-4. Available at SSRN: http://ssrn.com/abstract=1090368.
[5] Rochet and Tirole (2003), supra at fn 1.
[6] See Evans, D S. and Schmalensee, R (2007) Catalyst Code – The Strategies Behind the World’s Most Dynamic Companies, Massachusetts: Harvard Business School Press.
[7] Carlton, D W. and Perloff J M. (2005) Modern Industrial Organisation, 4th ed., London: Pearson Higher Education, page 82.
[8] Evans (2008), supra at fn 4, page 5.
[9] See Hagiu, A (2009) ‘Quantity Vs. Quality and Exclusion by Two-Sided Platforms’, Harvard Business School Strategy Unit Working Paper No. 09-094. Here, the author gives the example of video game console manufacturers such as Microsoft, Sony and Nintendo, which vet the quality of all game developers and exclude those not authorised to produce games on their platforms, by including security chips in the console. This is to give game players some level of quality assurance and to encourage them to buy both the console and its games with confidence.
[10] Supra at fn 9.
[11] Hagiu (2009), supra at fn 11, page 7.
[12] Boudreau K J. and Hagiu A (2008) ‘Platform Rules: Multi-Sided Platforms as Regulators’, Harvard Business School Strategy Unit Working Paper No. 09-061.
[13] To encourage consumption of these products, due to the external benefit accruing to the rest of society when an individual cures ailments.
[14]Evans, D S. and Schmalensee, R (2008) ‘Markets with Two-Sided Platforms’, Issues in Competition Law & Policy (ABA Section of Antitrust Law), Vol. 1, Ch. 28.
[15] Rochet, J C. and Tirole, J (2006) ‘Two-Sided Markets: A Progress Report’, RAND Journal of Economics, Vol. 37, No. 3, 645-667.
[16] Supra at fn 15, pages 664-5.
[17] Indeed a brief look back at Table 1 reveals that, for most of the platforms listed there, Side B customers get the product at or below cost, possibly even free. In contrast, most of those listed in Side A pay significantly above marginal cost.
[18] Evans and Schmalensee, supra at fn 14.
[19] Evans (2008) supra at fn 4.
[20] See Evans, D S. (2008) ‘How Catalysts Ignite: The Economics of Platform-Based Start-Ups’. Available at SSRN: http://ssrn.com/abstract=1279631.
[21] Supra at fn 4.
[22] Supra at fn 4, page 13.
[23] Due to the need for heavy research and development and sometimes very specific assets.
[24] Since much of the output in these industries consists of computer code rather than any physical good.
[25] Shapiro, C and Varian, H (1999) Information Rules, Massachusetts: Harvard Business School Press.
[26] This is in contrast to single-homing, where customers only use one platform – an example being video game players, who tend to choose either the X-Box or the PlayStation but rarely both, as this would be a duplication of cost with little or no added network benefit. Conversely, multi-homing occurs when using several platforms does enable the customer to tap into a wider market and exploit greater network benefits. Examples include home sellers who may list their property with several estate agents at the same time, to reach more potential buyers; equally, merchants may wish to accept both Visa and MasterCard to increase the pool of cardholders who wish to pay on credit. Another reason for multi-homing may simply be that platforms differentiate themselves and cater to different tastes and preferences, thereby encouraging customers to use more than one platform. See Rochet, J C. and Tirole, J (2006) ‘Two-Sided Markets: A Progress Report’, RAND Journal of Economics, Vol. 37, No. 3, 645-667.
[27] Evans (2008) supra at fn 4.