Wednesday, March 16, 2005

Michael: Policy Markets and Aggregate Decision-making

I was wrong.

When DARPA suggested using public futures trading markets to help analyze trends and political events in world politics, I was skeptical and condemnatory. Like many others I was shocked by the idea of betting on events at the Policy Analysis Market (PAM) such as assassinations and coups. This characterization was really a spectacularly bad example of the events and metrics such a market is best suited to track and predict.

There are good arguments on both sides of the issue as to whether such models are ethical and efficacious. What finally settled my hash on the pro side of predictive markets is reading James Surowiecki's "The Wisdom of Crowds". A chapter of his book deals with intelligence gathering in a strategic sense. He examines the pro and cons of centralization, decentralization, local control, and aggregation in cybernetic systems. Surowiecki discusses some of the strengths and weakness of markets in general, and makes some very good points about the ethics and efficacy of the proposed PAM, specifically.

In short, Surowiecki points out that we make speculations about devastating and ethically questionable occurrences every day as a routine, even mandatory, part of government. Why should it be morally questionable to ask such questions outside of government? At base, the purpose of such markets is to bring to bear a much more diverse and independent set of viewpoints than is possible inside of government. In theory, such markets could even draw intelligence in the form of price indicators out of our enemies. It is highly pejorative, and factually dubious, to characterize such markets as 'betting on assassination'. Their purpose and value lies in predicting the health and stability of foreign societies about which our policy elites know little, not in making speculations about low probability and secretive events such as assassinations.

NetExchange, the private company which developed the PAM in partnership with DARPA, proposed in 2004 to bring out a non-governmental version of the PAM that specifically excludes violent events from trading. To date, this has not yet happened, but I think it will be an interesting experiment when it does. Surowiecki's book is upsetting a lot of my preconceptions about the capacity of groups in different circumstances to make intelligent decisions. I recommend you give it a read if you are interested in social behavior.

0 Comments:

Post a Comment

<< Home




Feeds:
RSS/Atom Feed Site Meter
Powered by Blogger