COS/ECO 444: Electronic Auctions
Spring 2002
Prof. Ken Steiglitz,
ken@cs.princeton.edu.
Dept. of Computer Science, Room 421 CS Building
Teaching Assistant:
Ting Liu,
tliu@cs.princeton.edu
Dept. of Computer Science, Room 316 CS Building
Undergraduate Coordinator:
Tina McCoy,
tmmccoy@cs.princeton.edu
Schedule:
Lectures:
1:30-2:50 Tuesday and Thursday, Room 102 CS Building
Precepts:
Feb. 14, 28, and TBA, Thursday, 7:30 pm, Room 102 CS Building
Course Description:
The goal of this course is to connect theory to
real-world electronic auctions. Basic results will be derived and
tested
using both home-grown and commercial Internet auctions, bots, and humans.
Topics include: introduction to game theory, Vickrey auctions,
revenue equivalence, optimal auctions, multiple-unit auctions,
externalities, auction mechanism design and current internet auctions.
Outline of course:
Below is a tentative outline of what will be covered in each week and
the associated reading.
Note:
Background we need will filled in as needed throughout the term:
game theory, mathematical techniques, probability.
R. Gibbons, Game Theory for Applied Economists,
Princeton Univ. Press, Princeton, 1992, is on reserve.
Note:
Problem sets to provide practice with the theory will be
assigned throughout the term.
Week
- (2/5,7)
- Outline of course components, scope, organization, grading.
- In-class experiment: an auction.
-------------------------
- Brief history of auctions: Babylon, Rome's atrium auctionarium,
Didius Julianus,
fish auctions around the world, mail bid sales,
modern auction houses, internet auctions. R. Cassady, Jr.,
Auctions and Auctioneering, Univ. of Calif. Press, Berkeley, 1967,
is on reserve.
- In-class experiment: an auction.
- Review of some probability theory, order statistics.
- (2/12,14)
- Finish review of order statistics we'll be using.
(Some notes for this are posted on the course home page, see the
banner item Misc. Notes.)
- In-class experiment, another second-price auction.
- Vickrey (1961): types of auctions,
Vickrey Auctions, Symmetric Bayesian Nash Equilibria (SBNE),
dominant strategies, sincere bidding.
- Brief comments on
Assignment #1:
"Watching eBay". Goals of data collecting and description
of the first benchmark, Assignment #1A (download anything).
-------------------------
- Review experiment of 2/12
- Theory, continued: formalism for finding SBNEs.
- In-class experiment: expectations, a beauty contest.
- Shilling on eBay?
- Note: Precept: Thurs. Feb. 14, 7:30 pm,
Room CS 102.
Ting Liu will review some things that will help with Assignment #1A.
Click here for precept outline.
- (2/19,21)
- Graphical interpretation of the expected payment in second-price auctions.
- Theory continuing, Vickrey (1961):
Symmetric Bayesian Nash Equilibria (SBNE) in
first-price IPV auctions (simplest case, F= uniform, n=2), bid shading.
- A list of mathematical tricks for auction theory.
- derivative of inverse function
- Leibnitz's rule
- integration by parts
- integrating factor
- In-class auction experiment, third-price IPV auction.
-------------------------
- Discussion of third-price IPV auction experiment,
results, intuition, strategic analysis.
- Due Thurs., Feb. 21:
Assignment #1A, discussion of techniques.
- Theory continuing: SBNE for IPV first-price auctions
with arbitrary value distribution F, n bidders.
- In-class auction experiment.
- (2/26,28)
- News U can Use.
- Note: precept Thurs. 2/28, 7:30 pm, CS 102.
Topic: Assignment 1B, due Thurs. 3/7. Brief description
of what's required.
- Discussion of experiment of 2/21, non-uniform value distribution,
first-price, calculation of NE.
- In-class experiment, Dutch with recorded intentions.
- Theory continuing, Vickrey (1961): first-price and second-price
revenue equivalence.
-------------------------
- News, video clips if possible.
- Theory continuing, revenue non-equivalence with asymmetries,
no first-price/second-price ranking.
- Discussion of previous in-class experiment; in-class experiment.
- Beginning
Riley and Samuelson (1981).
- (3/5,7)
- Continuing Riley and Samuelson (1981): revenue equivalence.
- Discussion of previous in-class experiment
(jar of dimes), Winner's Curse.
- In-class experiment: Bidding for Paramount.
-------------------------
- Due Thurs. March 7:
Assignment #1B (download completed auction)
due, discussion of results, tracking auctions on eBay,
observed behavior on eBay.
- Comments and suggestions for Assignment #1C (extra credit).
- Continuing Riley and Samuelson (1981): optimal reserve prices.
- (3/12,14)
- Robinson's Problem, part of Problem Set #1 (Due Thursday, March 28).
- Continuing Riley and Samuelson (1981):
optimal reserve prices,
alternative auctions: Sad Loser.
- In-class auction.
-------------------------
- Continuing Riley and Samuelson (1981):
alternative auctions: Santa Claus, Matching.
- In-class auction.
- Discussion of term paper topics.
------ Midterm Break ------
- (3/26,28)
- News U can Use: DOJ probe of eBay
- In-class experiment, designing auction bots, auction tournament, Round 1.
- Review of the in-class auction of 3/14, downsloping density.
- Bulow and Roberts (1989):
How to run an optimal auction.
- Assignment #2:
running your own local auction. Introduction, get started.
- Due Tuesday, March 26:
Assignment #1C (any extra credit contributions?).
- Risk aversion and its effect on equilibrium bidding strategies.
-------------------------
- Risk aversion, continued.
- Relaxing assumptions: Remark about possible lack of efficiency
in optimal auctions (Bulow and Roberts). Also remark about lack of
revenue equivalence in asymmetric case (Vickrey) and with risk-averse
bidders (Riley and Samuelson).
- Discussion of Robinson's Problem
- In-class auction: first-price sealed bid with reserve.
- Due Thursday, March 28: Problem Set #1.
- (4/2,4)
- Briefly go over Problem Set #1. Vote on our recommendations to Robinson.
- Review of the in-class auction of 3/28,
first-price sealed bid with reserve.
- Speaking of term papers, here's an example of an empirical
market study: D. Lucking-Reiley,
"Field Experiments on the Effects
of Reserve Prices in Auctions: More Magic on the Internet",
now University of Arizona, version of 28 December 2000.
- Auction tournament, Round 2. (Same rules as Round 1; the bidding
strategies from Round 1 are posted here.)
-------------------------
- Discussion of Rounds 1 and 2 of the auction tournament.
Nash equlibrium strategy in a relative profit, first-price auction.
- Morgan (2001): "Efficiency and
demand reduction in multi-item auctions." See also
L. M. Ausubel and P. Cramton (1998):
"Demand reduction and inefficiency in multi-unit auctions."
- (4/9,11)
- In-class auction tournament, Round 3 (new conditions).
- So-called Dutch auctions on eBay: an example of a uniform price
multi-item auction. Is there an incentive for demand reduction? Is
it inefficient?
- Bulow and Klemperer (1996):
auctions vs. negotiations.
-------------------------
- Results of in-class tournament, Round 3.
- Bulow and Klemperer (1996):
concluded. Their result and how they spin it.
- Survey of some real auctions and their mechanisms, mail-bid
sales, buy-or-bid, local auctions, choosing mechanisms for different
applications.
- Note: Precept: Thurs. April 11, 7:30 pm,
Room CS 102. Topic: Assignment #2.
- Due: Thurs. April 11:
Reports from student on their choices of term paper topics.
- (4/16,18)
- In-class auction tournament, Round 4 (same conditions as Round 3).
Since you have time to work on this, and some of you now have simulators,
I've decided to raise the bar a little. Ting and I will submit two
bidding functions in addition to yours, and the payoff will be:
$20.00 for first place, $15.00 for second, and $10.00 for third,
with ties resolved in your favor.
- Myerson and Satterthwaite (1983):
The Revelation Principle, and their Impossibility Theorem.
- Shilling: definition, ethics thereof, discussion.
-------------------------
- Handing out Problem Set #2 (Due Tuesday, April 30).
- Review of Round 4. Since you people haven't gotten it,
we'll go for Round 5, same conditions as Rounds 3 and 4,
next Tuesday.
- The many ways eBay differs from Vickrey auctions;
fakes and frauds on eBay.
- An agent-based simulation of eBay: snipers vs. ducks, based on
H. Mizuta and K. Steiglitz,
Agent-based simulation of dynamic online auctions,
Winter Simulation Conference, Orlando, FL, Dec. 10-13, 2000.
- Double auction markets, introduction, agent-based simulation as a tool.
- (4/23,25)
- In-class auction tournament, Round 5 (same conditions as Rounds
3 and 4). Are people learning, computing, or guessing the Nash
equilibrium?
- Simulation techniques (con't): General discussion of simulation
as a tool: pluses and minuses. Simulating double-auction asset markets,
stabilizing effects of value traders, destabilizing effects of
trend-chasers, extreme sensitivity of markets to information availability,
price bubbles.
- Bulow and Klemperer (1994):
rational frenzies and crashes, studying market behavior using auction theory.
-------------------------
- Due: Thursday, April 25:
Assignment #2, discussion of results. (You can also try to sell
some things for the rest of the semester if you'd like.)
- Continuing with Bulow and Klemperer
(1994): Simulation of Bulow and Klemperer's model.
- In-class experiment (the tournament is over, this is something
a little different).
- Mock auctions, running ("trotting"), Buyer rings
("kippers", "pies"), knock-out sales (the London antiquarian book ring
and Leo Marks). Anyone read Leo Marks, "Between Silk and Cyanide:
A Codemaker's War, 1941-1945," Harper Collins, 1998?
- (4/30,5/2)
- Due Tuesday, April 30: Problem Set #2.
I'll go over solutions.
- Graham and Marshall (1987):
a theory of buyer collusion.
-------------------------
- Competition and spite: "The Spite Motive and Equilibrium Behavior
in Auctions," J. Morgan, K. Steiglitz, and G. Reis, working paper,
version February 23, 2002; (pdf)
- Further discusssion of tournament results.
Evolutionarily Stable Strategies (ESS), Nash equilibria,
replicator dynamics, and spiteful behavior.
- Postmortem: your constructive or creative suggestions?
------------------------------------------------------------
End of Reading Period, May 10, 13 and 15:
Presentation of student papers.
From the sixties.
Left: Peter Chance auctioning a van Gogh portrait at Christie's
in London; Right: Sale of cut flowers by Dutch clock in Aalsmeer, Holland
(both from R. Cassady, Jr., Auctions and Auctioneering,
University of California Press, Berkeley, 1967). See also the current
v.B.A. in Aalsmeer in Holland (the Bakkie Koffie is fresh all day,
ask for Hanns), and some
Canadian Dutch-auction flower markets.
I owe a great debt to
Prof. John Morgan,
rjmorgan@princeton.edu,
who co-taught the first
instantiation of this course, and shaped it in many ways.