Using randomized selling schemes for vertical differentiation on the Internet.

摘要

This dissertation investigates price discrimination that involves selling through two channels that differ in their risk attributes. In one channel prices and qualities are fixed and availability is assured. In the second channel, the seller offers a distribution of prices and qualities and may not guarantee availability. The dissertation uses two related models to offer explanations for observed behavior of online sellers and to discuss implementation issues in recent e-commerce environments.;The first model is used to characterize an optimal scheme for the sale of multiple identical items by a monopolist in a market comprising risk-averse buyers. It establishes that a seller can obtain segmentation benefits by randomizing prices in one channel while also offering a risk-free alternative in another. The optimal vehicle of such randomization is a draw from a discrete two-point probability distribution function.;A second model is used to characterize optimal selling strategies for a monopolist who sells two products of distinct qualities. It shows that it can be optimal to offer the high-quality product alone in a random sale event. It can also be optimal to offer the low-quality product alone in a sale. However, the seller cannot benefit by offering both high and low-quality products in a sale that is held with probability less than one.