Exclusive dealing with network effects

Abstract

This paper explores the ability of an incumbent to use introductory offers to dominate a market in the face of a more efficient rival when network effects rather than scale economies are present. Both in the case of one-sided and two-sided markets, for introductory offers to be profitable when consumers can multihome, they need to be discriminatory and exclusive. In this setting, exclusivity as opposed to just commitment to purchase is critical — consumers must commit not to purchase from the rival in the future in order that introductory offers can work. The use of such contracts is anticompetitive and inefficient but does not necessarily result in complete foreclosure.

Introduction

In existing models of naked exclusion such as in Rasmusen et al. (1991) and Segal and Whinston (2000), scale economies allow an incumbent to exclude a rival by signing up customers to deny the rival the necessary scale to profitably enter. This paper develops a framework where naked exclusion instead takes place through network effects. The incumbent can sign up consumers before a rival firm offering a product with a superior network reaches the market. However, this by itself is not sufficient for the incumbent to make sales since consumers will wait to buy from their preferred firm (which will anyway enter given there are no scale economies). A necessary condition for the incumbent to profit from introductory offers is the ability for these to be offered in a discriminatory way. Attracting some consumers with low prices raises the network benefits the incumbent can offer to other consumers, which, in turn, allows it to charge more when it competes with the entrant. However, each signed-up customer reduces the size of the customer base that the incumbent can later exploit. The incumbent therefore optimally designs an introductory offer which balances these opposing effects.⁴

This simple logic, however, does not hold once we take into account the natural tendency for consumers in network markets to want to multihome (that is, buy from both firms so as to enjoy greater network benefits). When consumers can multihome, the incumbent may no longer be able to profit just by signing up consumers before competing with the entrant. Provided it is not too costly to do so, consumers that sign with the incumbent will also buy from the entrant to take advantage of its superior network.

In order that signing up consumers in the initial stage remains profitable for the incumbent, it must instead do so exclusively. Consumers that sign with the incumbent not only have to commit to buy from the incumbent but they must also agree not to purchase from the entrant later. Where such contracts are feasible, the incumbent will profitably sign up some of the available consumers, doing so exclusively, and then exploit those that do not sign subsequently. Such exclusive deals raise the incumbent's profit at the expense of the entrant, at the expense of those consumers not offered exclusive deals and at the expense of efficiency.

Interestingly, such exclusive contracts may not involve complete foreclosure. In a one-sided market setting, consumers that do not sign exclusive contracts may multihome, buying from the incumbent to reach signed-up consumers and buying from the entrant to take advantage of its superior network. Such multihoming results in dampened ex-post competition, which also benefits the entrant to a certain extent. This, along with the fact that signed-up consumers do not join the entrant's superior network, is a source of potential harm. In fact, we show a ban on the use of exclusive deals raises consumer surplus and welfare. Nevertheless, given the presence of multiple firms in the competition stage, unlike the traditional naked exclusion results, the adverse effects of exclusive deals may be more difficult for a competition authority to detect in our setting.

Our main conclusions also carry over to a two-sided market setting, where the incumbent signs up “sellers” exclusively and then exploits “buyers” who want to interact with these sellers. Here discrimination is natural given there are two different types of consumers and given that it may only be feasible to get one side of the market to sign exclusive deals (e.g. if buyers are households and sellers are firms). Buyers have all their surplus exploited while sellers benefit from exclusive deals. In our model with two-sided markets, exclusive deals lead to complete foreclosure, reflecting that all sellers sign exclusive deals so there is no reason for buyers to multihome. Moreover, sellers may be the strongest supporters of exclusive deals in this setting.

Throughout the paper we assume consumer expectations that minimize the scope for coordination failures amongst consumers. For instance, the coordination failure in the existing naked exclusion literature, in which consumers signing contracts would do better if they could coordinate on an equilibrium that allows for entry, does not arise in our setting. Despite this, consumers in aggregate are still worse off as a result of exclusive deals. In a one-sided setting, this reflects the ability of the incumbent to divide consumers' interests, by making an attractive offer to a limited number of users initially, and then exploiting the remaining users in the subsequent competition stage. In two-sided markets, this strategy arises even more naturally, due to the inherent segmentation of consumers into two groups.

The frequent use of exclusive contracts in markets with network effects has been documented by Balto (1999) and Shapiro (1999). They provide detailed discussions of exclusionary actions in a number of these markets, including ATM networks, computer reservation systems, credit card networks, floral delivery networks, Pay-TV, video game systems and the market for wire money transfers. Here we briefly note one such example, which suggests exclusive dealing may play an anticompetitive role in network industries.

The Florist Telegraph Delivery (FTD) network developed in the 1940s so that people could send flowers to distant locations through other member florists. It enjoyed the participation of the majority of large florists. To prevent florists trying to develop their own networks, FTD adopted an exclusionary rule that florists could only be members if they did not belong to any other such network. In the face of strong network effects, these exclusive membership rules made it difficult for new networks to offer their customers a comparable service. It was not until the Antitrust Division and the FTD entered into a consent decree in 1956, in which the FTD dropped its exclusive membership rule, did other floral networks such as AFS and Teleflora emerge. This decree still remains in effect — in fact, in 1995 it resulted in an Antitrust Division action over FTD's incentive program “FTD only” which was also deemed to have a similar exclusivity effect.

The rest of the paper proceeds as follows. Section 2 provides a brief discussion of how our paper relates to the existing literature. Section 3 develops our basic framework. Subsequent sections consider the case of a one-sided market where multihoming is not feasible (Section 4), the case of a one-sided market where multihoming is feasible (Section 5), and the case of a two-sided market where multihoming is feasible (Section 6). We conclude in Section 7.