Econ 521b: Dynamic Mechanism Design (2012)

9:00-10:20 M W, 28 Hillhouse, Room 106

Dirk Bergemann
30 Hillhouse Avenue, #24
432-3592
Email
Office hours T 1:00-3:00

Juuso Valimaki
30 Hillhouse Avenue, #6
432-6519
Email
Office hours T 1:00-3:00

Topic: This course covers selected topics in the economics of information and uncertainty. The theme of the course this year is “Dynamic Mechanism Design”.

The theory of mechanism design as laid out in the seminal contributions of Vickrey (1961) and Hurwicz (1972) analyzes the performance of various institutions in an economy where information is dispersed amongst participating agents. The key underlying idea is that information is similar to other economic goods in the sense that the agents have autonomy over the decisions regarding their information. How information is used depends on the mechanism in place. The first generation of (positive) results for mechanism design in Clarke (1971), Groves (1973), Green and Laffont (1977) and D’Aspremont and Gerard-Varet (1979) characterized surplus maximizing mechanisms within private values economies. The second step towards a general theory resulted from a change in focus. Rather than start with surplus maximization as the objective, the work of Myerson (1979), Myerson (1981) and Myerson and Satterthwaite (1983) on revenue maximizing mechanisms started with a characterization of all implementable allocation rules. This opened up the field for a large number of applications ranging from optimal auctions and bilateral trade to regulation and negotiations. Perhaps surprisingly, almost all of the theory and most of the applications are set in a completely static economic model. The available information comes in the form of an initial endowment and it is collected from the agents in a static revelation game. Furthermore, the allocation is decided once and for all. If information on the potential uses of a resource arrives over time, then it is quite likely that any optimal allocation mechanism also reallocates the resource over time. The recent literature on dynamic mechanism design has started an investigation of surplus maximization (for example in Bergemann and Välimäki (2010) and Athey and Segal (2007)) and revenue maximization (e.g. Pavan, Segal, and Toikka (2009)) in models that incorporate these dynamic features. The objective of this course is twofold: First, it is meant to provide an introduction into the main themes and techniques of dynamic mechanism design. Second, it is meant to illustrate the central insights of the theoretical work in the context of important economic applications.

Course Requirements: This course has five basic requirements. They are: (i) reading the assigned papers before the presentation in class, (ii) solving the problem sets, (iii) presenting one research paper, (iv) writing four referee reports (before the presentation of the papers) and (v) writing a term paper. The assignments will be given biweekly. Class participation, assignments, and the term paper will jointly determine the final grade. The due date for the term paper is Monday, May 14th, 4pm in the mailbox of Professor Dirk Bergemann.

Readings: The current reading list will be completed as the class proceeds.

Syllabus: The complete syllabus for Econ 521b.

Weekly Schedule::

1. 1/9 – 1/16 (JV): Efficient Mechanisms: Vickrey (1961), dAspremont and Gerard-Varet (1979), Cremer and McLean (1988), Jehiel and Moldovanu (2001), Neeman (2004)

Lecture1 , Lecture 2, Readings for Static Mechanism Design, Bergemann and Valimaki (2010), Bergemann and Valimaki (2006)

2. 1/18 – 1/25 (JV): Efficient Dynamic Allocation: Bergemann and Valimaki (2006), Bergemann and Valimaki (2010): Dynamic Pivot Mechanism: Slides, Gershkov and Moldovanu (2009), Gershkov and Moldovanu (2011), Gittins Index Theorem

Problem Set 1

Referee Report: How to Write a Referee Report; (i) Rahman, (ii) Goeree and Kushnir (iii) Pai and Vohra

3. 1/30 – 2/1 (DB): Dynamic Incentives and the Ratchet Effect: Baron and Besanko (1984),

Problem Set 2, Sequential Screening

4. 2/6 – 2/8 (DB): Arrival of New Information: Courty and Li (2000), Eso and Szentes (2007), New Information

5. 2/13 – 2/15 (DB):

(i) Dynamic Moral Hazard: Bergemann and Hege (2005); Horner and Samuelson (2012), Mason and Valimaki (2011), Bonatti and Horner (2011), Deb (2011)

(ii) Information, Information Disclosure and Mechanism Design: Bergemann and Pesendorfer (2005), Bergemann and Morris (2011), Kamenica and Gentzkow (2011), Rayo and Segal (2011)

Information and Mechanism Design

6. 2/20 – 2/22 (DB): Guest Speaker: Tim Roughgarden, Stanford University

Approximation in Algorithm Game Theory: The Price of Anarchy

7. 2/27 – 2/29 (DB): Infinite Time Horizon Models Pavan, Segal, and Toikka (2009), Guest Speaker: Marco Battaglini, Princeton University