Essays on new Internet market places

Abstract

This thesis uses the techniques of economic theory to examine the behaviour of agents in new marketplaces that have formed on the Internet. It is divided into three chapters, each focusing upon a different aspect of market behaviour. I begin in chapter 2 by building a simultaneous, unit demand, heterogeneous good, ascending auction model, which I use to study the behaviour of bidders in Internet auctions. Using this model, I demonstrate that late bidding—often observed in eBay auctions—is not inconsistent with allocative efficiency. Moreover, I show that when all agents are fully rational, have private values, and face a perfect bid-transmission mechanism, there can still exit an incentive for bidders to systematically delay their final bid. In a manner consistent with earlier empirical observations, this incentive disappears when the hard-close ending rule is relaxed.
In chapter 3 I model strategic interaction amongst search engines that compete to serve consumer needs. Search engines generate revenue from advertisers, but also provide free organic search results. I demonstrate that, in an attempt to win market share, search engines compete not only against each other, but also against themselves: providing high-quality free links that compete for clicks with their own advertisements—thus cannibalising their advertising revenues. In particular, I find that in equilibrium consumers always (at least weakly) prefer to click on at least one non-paid-for link before clicking on a revenue generating advertisement so that some consumers never click an ad at all. That notwithstanding, revenue cannibalisation provides an incentive for quality degradation, even when the provision of quality is costless, and may engender low quality equilibria. When search engines show differentiated advertisements, the incentive to reduce quality is particularly strong.
Chapter 4 examines the relationship between the transmittability of information via advertisements and the fee structure used by the advertisement’s publisher. For an advertiser, sending general advertisements with inflationary claims may attract additional consumers with whom it is poorly matched. This is costly for the firm when it must pay for the ads on a per-click basis (i.e. when it must pay for each consumer visit that it receives) since many of its visitors will not purchase. As a consequence, I find that perfect information transmission can always be sustained when adverts are priced per-click. By contrast, when firms pay for advertisements on a per-impression basis or on a per-sale basis, there is no disincentive to attracting poorly matched consumers, and maximum profits are obtained by attracting all consumers with some positive probability of purchase. This feature undermines the existence of fully informative equilibria under such fee structures, and may result in no information transmission being possible at all. Consumers benefit from increased informativeness, but distortions introduced by the market power given to advertisers imply that society may be better-off with no information transmission taking place.

More information

Identifiers

Catalogue record

Export record