The Dr. Strangelove of the Information Age, Vice Admiral John M. Poindexter (ret.), is being forced out as director of the Department of Defense’s Information Awareness Office (IAO), a unit of the Defense Advanced Research Projects Agency (DARPA). The departure comes in response to the public outrage over his proposal to establish an online stock market in terrorism futures, formally known as the Policy Analysis Market (PAM). (A copy of the PAM site is here.)
This stillbirth follows closely upon the early demise of his no less widely disparaged Total Information Awareness (TIA) plan. In that scheme, the federal government would mine databases of every consumer’s credit-card purchases, car rentals, library books, airline-ticket purchases, telephone bills, official documents, and the like, ultimately in real time, looking for patterns that might indicate terrorist planning or activity.
This is the same John Poindexter who, as National Security Advisor to Ronald Reagan, was convicted of deceiving Congress in the Iran-Contra cover-up and was sentenced to six months in prison for it.
In Poindexter’s PAM plan, investors express their predictions about future political and economic events such as assassinations, bombings, or missile attacks in the Middle East by buying and selling futures contracts, just like traders buy and sell futures contracts on oil, pork bellies, coffee, and so forth on commodities markets. Investors who think a political figure or a government will be assassinated or overthrown would buy a futures contract; if one thinks the murder or coup will not occur, one could sell a contract. The Economist Intelligence Unit and Net Exchange, a derivatives software company, helped devise the DARPA-funded marketplace.
Politicians pushed their way to the nearest bully pulpit and editorial writers sharpened their pencils to denounce the idea as “grotesque,” “ridiculous,” “morally wrong,” and “bizarre.” Maybe so, but it’s not dumb. Defended by writers at both the liberal New York Times (July 31, 2003) and the conservative Wall Street Journal (August 1, 2003), information markets like PAM are likely to be one of the Internet’s important contributions.
All markets aggregate and communicate information dispersed among multiple knowledgeable participants. They are most useful when a lot of people each have a little bit of information, because every market creates value by aggregating things that don’t mean much on their own. When put together, these scraps of information become something meaningful. In the case of a futures market: a constantly evolving collective forecast.
Contrary to impressions in the media, the primary purpose of PAM was not to predict individual events of terrorism, but rather to predict inputs into terrorism, such as the economic growth rates, political instability, and military activity in countries of the Middle East. Moreover, the most likely version restricted participation to invited experts from government, academia, and industry to develop expert consensus forecasts on war- and terror-related topics.
In addition, using new combinatorial methods devised by George Mason University professor and decision-market champion Robin Hanson, PAM’s trading process allowed traders to structure combinations of futures contracts to represent predictions about interrelated issues. The Pentagon would post the original odds on each outcome, and the market could calculate the literally millions of possible combination scenarios. What would happen to unrest in the Middle East if the U.S. withdrew its troops in Saudi Arabia? How would Jordan fare politically if the “road map” were successfully implemented?
Headline-hungry politicians and an accommodating sensationalist press presented a macabre vision of a death-and-destruction bazaar. In reality, PAM was a lot less scary and a lot more important as a good example of an increasingly vital information application. Historically, most markets have been created to make money with information as a byproduct. In contrast, these markets are being created specifically for the express purpose of ferreting out and organizing information.
The public relations fiasco will surely dampen open enthusiasm for these markets. Fortunately, the potential contributions of information markets to decision making remains undiminished, and they’re widely applicable — including for developing marketing strategy.
Any organization that can frame its informational needs as predictions can set up an investing game where the trading patterns of experts, frontline personnel, or customers will generate streams of highly condensed, real-time forecasting data. Since the Internet makes such markets so quick and easy to set up, they could still proliferate. In fact, several first-generation information markets are up and running on the Web today. We will survey these early efforts next time.