TUTORIAL: Markets in uncertainty: Risk,
gambling, and information aggregation
This
tutorial will survey the economic and computational aspects of markets in
uncertainty: financial instruments that pay off based on the realization of an
uncertain variable.
Markets
in uncertainty effectively allow traders to place bets on the future outcome of
an uncertain proposition or variable.
Examples include stock markets (like NYSE or NASDAQ), options markets
(like CBOE), futures markets (like CME), other derivatives markets, insurance markets,
and sports betting markets, and have been called in various contexts securities
markets, information markets, prediction markets, or contingent claims.
Although historically such markets have earned a somewhat mixed reputation, in
truth they serve two very important economic and social functions. First, they
help people manage risk by allowing traders to hedge, or to insure against
undesirable outcomes. For example, the owner of a house may purchase insurance
to hedge against unforeseen damage to the house. Or the owner of a stock might buy
a put option to insure against a stock downturn. Second, markets in uncertainty
help aggregate and disseminate information, by giving traders the incentive to
speculate, or to trade when market prices do not reflect their assessment of
the likelihood of future outcomes. For example, a trader might buy soybean
futures if he suspects a bad crop will drive prices up, regardless of risk
exposure. Or a gambler might bet on a football team if his assessment of the
team's chances of winning is greater than what the going odds reflect, modulo
fees. According to economic theory, when many traders with different information
all speculate, the equilibrium price reflects the sum total of all of their
information. Much supporting evidence can be found in empirical studies of real
markets and laboratory experiments. Moreover, in most markets, prices are
publicly available, and so provide significant value as a mechanism for
disseminating information to the masses.
The
tutorial will cover the essential economic background. We will start at the
agent level, introducing subjective probability, utility maximization, and
risk. We will then move on to the mechanism level, covering classical results
regarding Arrow-Debreu securities markets and
rational expectations equilibrium theory, and more recent empirical studies of
options markets, political stock markets, sports betting markets, horse racing
markets, and market games, and laboratory investigations of experimental
markets. The tutorial will also survey recent research on the computational
aspects of these markets, including new market concepts called compact markets,
compound markets, combinatorial markets (yes, they're all distinct!), and
distributed
computation in markets. We will also touch on the legality of markets in
uncertainty, and current efforts to field new markets in the face of legal and
regulatory obstacles.
Biosketches of the
presenters:
David
M. Pennock is a Computer Research Scientist at Overture
Services Inc., in
Web.
His research has received significant attention among e-commerce companies and
in the media, including reports in Discover Magazine, New Scientist Magazine,
and the New York Times. For more information, please visit http://dpennock.com/
Michael
P. Wellman is a Professor of Computer Science and Engineering at the