This paper, which can be found here, It aims to clarify whether and to what extent two-sided platforms are different from platforms selling complementary products. It also seeks to explain why the distinction matters for the purposes of competition assessments of firms’ behaviour.
The paper is structured as follows:
A first section explains why firms operating in two-sided markets are different from firms selling complementary products.
According to the economic literature, a two-sided platform is a firm that sells two different products or services to two groups of consumers, where the demand from one group of consumers depends on the demand from the other group and, potentially, vice versa. In other words, demand is affected by indirect network effects (i.e. consumers’ willingness to pay for a product depends on the number of consumers (or the quantity bought) of another product). A platform internalizes these indirect network effects.
There are differences between platforms and firms selling complements. A first difference is that, in the case of two-sided platforms, one (albeit only one) of the links between the demands may be negative. In other words, demand from one customer group may decline with higher sales to the other group of customer (e.g. excessive advertising may lead to less use of media). A second and more important difference is that, according to the definition of a two-sided platform, the buyers of the two products do not internalise the links between the two demands, which are therefore called externalities. Herein lies the crucial difference between a two-sided platform and a firm selling complement products: the two products sold by a two-sided platform are bought by different customers, unlike complementary products that are bought by the same customer. It is exactly because each customer buys only one of the two products sold by the platform that buyers typically are unable to internalise indirect network effects, and why a platform is required to this end.
A second part of this section discusses how different types of two-sided markets have different implications for antitrust analyses.
An important discussion in the US Supreme Court’s American Express v Ohio judgment (which I reviewed here) revolved around the distinction between two-sided and complementary markets, where the judgment cited an earlier article by the author (Filistrucchi, Geradin, Van Damme, & Affeldt, “Market Definition in Two-Sided Markets: Theory and Practice,” 10 J. Competition L. & Econ. 293, 297 (2014)).
As recognised by the U.S. Supreme Court, in cases involving two-sided platforms the distinction between two-sided transaction and non-transaction platforms is crucial. This distinction is important because it highlights a fundamental difference in the pricing strategies available to platforms and, therefore, in the way these firms compete. Two-sided non-transaction platforms (e.g. newspapers) are characterised by the absence of a transaction between the two sides of the market. Even though an interaction is present between the two sides of the platform, it is usually not observable by the platform, so that the platform is unable to set a per-transaction or per interaction fee. On the other hand, two-sided transaction platforms (e.g. payment cards) are instead characterised by the presence and observability of a transaction between the two groups of platform users. As a result, the platform is not only able to charge a price for joining the platform but also for using it, i.e. the platform can charge a two-part tariff.
While two-sided non-transaction markets are characterised by membership externalities (or indirect network effects), two-sided transaction markets are also characterized by usage externalities. Membership externalities arise from joining the platform (buying a newspaper or placing an ad in a newspaper, holding a payment card or having a point-of-sale terminal, listing your product at an auction or attending an auction), while usage externalities arise from using the platform (paying or accepting payment with a card, selling and buying a product at an auction). When a transaction is needed to use the services of a platform, one member of each customer group needs to agree with one member of the other group in order to use such services. It follows that the platform cannot sell its usage services unless both customers agree to buy them.
Justice Breyer, in his dissenting opinion, was correct to observe that in the case of payment card companies, which are two-sided transaction platforms, “the services resemble complements because they must be used together for either to have value.” Still, the products are not complements in an economic textbook sense because they are not bought by the same customers. It is true that demand for the products on one side may decline with a rise in the price on the other side. Yet, typically, there is no individual customer that finds the two products complementary because no customer wants to consume both. This has implications regarding price distribution, which will change the mix of customers on the sides of the market – an effect that does not exist as regards complementary products.
A third section explains why two-sided platforms resemble platforms selling complementary products.
Two-sided platforms take into account (or internalise) the network effects between the demands they face. By changing the price it charges to one group of customers, a platform will influence also the demand from the other group of customers, even if it holds constant the price charged to the latter group. When demand for the product sold to one customer group declines (or rises) with an increase (or decrease) in the price charged to the other group, the situation will be somewhat similar to that of complementary products. A second consequence of the internalisation of the network effects is that, in certain circumstances, competing two-sided platforms selling substitute products may behave (e.g. price) as firms in one-sided markets that sell complementary products. In general, it is the sign and the size of own- and cross-network effects, together with their internalisation by the platforms that determines whether the platform behaves as if it sold complementary products.
Yet, even when firms in two-sided markets behave as firms selling complementary products, the welfare outcomes of platform conduct may be very different. The reason lies, once again, in the fact that customers on the two sides are different and that, as a result, the two consumers’ welfare may not move in the same direction. Consider for instance the traditional complementary product case of an inkjet printer where each customer needs to buy one printer and 10 cartridges. Suppose the price of the printer declines by 10 dollars and the price of each cartridge increases by 1 dollar each, which means that the price and welfare of the customers will not change. Now consider a platform like pay-tv with advertising. 100 viewers buy one match subscription and there is only one advertising slot at half time that can be sold to a unique advertiser. Suppose that the price paid by the viewers declines by 10 dollars and the price paid by the advertisers increases by 1000 dollars. Again, the overall effect will be the same, but viewers and advertisers will feel it differently. On the other hand, advertisers will benefit from the likely increase in the number of viewers, while viewers may also enjoy a reduction in the length of the advertising due to the increase in the advertising price. In this case, the viewers’ welfare is likely to increase, while advertisers’ welfare may rise or decline.
A fourth section explains why the difference between two-sided platforms and selling complementary products matters for competition law.
Unlike with complementary goods, platforms can make consumer welfare move in opposite directions. This means that, in the context of platforms, competition policy needs to take a stance on which customer group, if any, should be given more consideration. This is not an issue in markets for complement products, in which there is only one customer and, hence, consumer welfare is either enhanced or prejudiced.
As a result, is not correct to say that “two-sided markets are like markets for complementary products.” More precisely, it is not correct except in very special circumstances, when referring to welfare effects of firms’ (pricing) strategies.
As I noted in my comment to the American Express v Ohio judgment, I found the dissent’s ‘assertion that platform markets are like markets for complements a bit odd – it may work as an analogy and there may be similarities between the two, but multi-sided markets are different creatures from markets with complementary products.’ My intuition when saying this is outlined in this piece: complementary products have effects on a single set of customers, while two-sided markets have (different) impacts on different sets of customers.
This distinction is important because it raises an important policy issue for competition law: which sets of consumers should competition law protect? The majority’s opinion in AmEx seems to think it does not matter what the impact on individual groups of customers might be as long as overall output increases. This is a normative choice, however, and we ignore it at our peril – and at the risk of divergence. Contra the U.S. Supreme Court decision, one of the characteristics of the payment card judgments in the UK was that they implicitly accepted that each set of customers would have to enjoy the benefits of a prima facie restriction of competition under Art. 101(3) TFEU – as is also argued in the article reviewed above. I expect this issue to raise its head more insistently in coming years.