Information and Learning in Mechanism Design
Abstract
This dissertation studies the design of mecanisms in settings where information acquisition or communication are significant features of the environment.The first chapter, coauthored with Xiaosheng Mu, studies a dynamic pricing model where buyers have the ability to learn about their value for a product over time. A seller commits to a pricing strategy, while buyers arrive exogenously and decide when to make a one-time purchase. The seller seeks to maximize profits against the worst-case information arrival processes. We show that a constant price path delivers the optimal profit, which is also the optimal profit in an environment where buyers cannot delay.
The second chapter develops a model of costly information acquisition, focusing on an application to scientific research. It shows that non-transparency can induce a scientist to undertake a costlier but more informative experiment if it also enables her to commit to acting scrupulously. Using this insight, this chapter demonstrates the general existence of non-degenerate experiment costs such that greater transparency in scientific methodology results in research choices that are worse for those interested in the results.
The third chapter develops a model of delegated project choice with mulitple agents, considering the impact of competition in these settings. Under an alignment assumption (which ensures the optimality of full discretion in the single-agent case), optimal mechanisms entail stochastic agent choice but deterministic project choice. Without alignment, deterministic project choice may be suboptimal. Without the ability to randomize allocations, competition can be harmful for the principal.
Terms of Use
This article is made available under the terms and conditions applicable to Other Posted Material, as set forth at http://nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAACitable link to this page
http://nrs.harvard.edu/urn-3:HUL.InstRepos:40050071
Collections
- FAS Theses and Dissertations [6566]
Contact administrator regarding this item (to report mistakes or request changes)