Essays on information and competition.

This dissertation aims to shed some light on a variety of topics regarding information and competition. The first three papers focus on the impact of "networks" (in a broad sense) as an information acquisition tool. Recent ICT-developments are causing consumers to become increasingly connected. The first two papers, therefore, considered the impact of (actual) social networks. By creating and accessing such networks, individuals can pool their information and as such become better informed. The first paper investigated how such networks may affect market outcomes when consumers search sequentially. The second paper considered the informativeness of networks in a more general setting. The network framework used in these two papers may also be interpreted more broadly. In a third paper, we expanded the framework and reinterpreted it as a model of passive information acquisition. The last paper focused on a topic that has intrigued economist for a long time (see Stigler, 1964): How can firms overcome the difficulties associated with the implementation of "collusive" agreements? In particular, we investigated the role a common supplier could play in disciplining a (horizontal) cartel by refusing to supply a defector. The main insights provided by this dissertation are summarized below. The first paper studied the implications of an information network in the presence of sequentially searching consumers. Previous work (see Galeotti, 2010) finds that networks can decrease consumers' search efforts, reducing the overall level of consumer informedness. Firms respond to this uninformedness by competing less aggressively and increasing prices. We showed that this "anti-competitive" effect of networks depends crucially on the mode of search and the timing of network activities. With respect to the mode of search, we showed that when consumers search sequentially and the market is duopolistic, larger networks always decrease the price paid by all consumers. When there are more than two firms in the market, larger networks may increase prices. With respect to timing, we showed that consumers can only become less informed if search takes place prior to using the information network. The second paper examined the informativeness of consumer information networks in a more general setting. We characterized the informativeness by deriving a simple formula describing how many different pieces of information a network is likely to contain. Given this characterization, we showed how the informativeness varies in the network size and the number of objects. We found that as the network grows larger, or the number of objects increases, consumers will become better informed. The informativeness of the network has important consequences for the degree of competition between firms. We showed that larger networks imply better information transmission, which intensifies competition and decreases prices. We also showed that in the presence of consumer information networks, an increase in the number of firms always decreases prices. As an extension, we also considered the impact of endogenous networks, star-shaped networks, costly sequential search, and network segregation. For many products, consumers acquire purchase-relevant information in a "passive" way. They go about their everyday life, and in the meantime they encounter information that is potentially relevant for a later purchase. When they eventually need to buy the product, they will recall part of this information and do not have to search "actively" for it. The third paper developed and estimated an empirical model of passive information acquisition. Passive information acquisition was modelled as a stochastic process in which consumers have opportunities of becoming informed about firm prices. We estimated our model using data on the Belgian market for diesel fuel. Our results indicated that most automobile drivers have few information opportunities and compare prices of a very small number of filling stations. A small fraction of drivers is, however, extremely price-comparing and has a large number of opportunities. We used our estimates to assess the effects of market structure on competition. We found that an increase in the number of filling stations decreases the expected price for all drivers by inducing them to become better informed. The fourth paper investigated the role a common upstream supplier (the "hub") can play in facilitating collusion between his downstream buyers (the "spokes"). We investigated a strategy in which the hub no longer supplies a defector in the downstream cartel. We showed that this strategy facilitates collusion by supporting collusion for lower discount factors. However, the cost of refusing to supply is borne by both the defector and the supplier. For the latter to participate, a number of conditions need to be fulfilled. We showed that only in a very limited number of cases, the supplier will participate in the punishment scheme. Only when the buyers are involved in strong price competition among themselves, and they have sufficient bargaining power vis-à-vis the seller, is the seller willing to impose an embargo to stabilize the downstream cartel.nrpages: 173status: publishe