The Dawn of Information Markets, Part 2

The stillbirth of Vice Admiral John M. Poindexter’s proposal for a market in terrorism futures could have a chilling effect on one of the Internet’s most important potentials — the use of information markets to aggregate and codify distributed intelligence.

Historically, most markets have been created to make money with information as a byproduct. In large part enabled by the Internet, markets in recent years have been created for the express purpose of ferreting out and organizing information.

The best known is the Iowa Electronic Markets at the University of Iowa, especially its Presidential Election Vote-Share Market. Since its inception in 1988, this pre-Internet information market has successfully predicted the outcome of every U.S. presidential election. It’s more accurate than traditional polling techniques in predicting the outcome of U.S. and foreign elections, primaries, and other political events. In fact, in tight elections, professional bond traders, who often have millions of dollars riding on post-election economic policies, monitor the Iowa political stock market as a leading indicator.

In pop culture, Hollywood Stock Exchange (HSX) is the best known. Movie fans speculate on box-office returns, opening-weekend performances, and the Oscars. The collective predictions by this enthusiast community about opening-weekend returns have proven more accurate than the forecasts of the movie industry itself. And, not surprisingly, the HSX last year correctly picked 35 of the 40 Oscar nominees in the top eight categories.

It’s not just for fun. HSX was founded by a spin-off of Cantor Fitzgerald, the global fixed-income trading firm. It knows how to make money with data; it sells HSX results to the movie industry.

From the left of the national political spectrum, the American Action Market opens this fall for trading in futures contracts relating to what the site claims are the two most important questions facing the world today: What will the U.S. government do next? What is informing the U.S. government’s current behavior?

Some contracts will be based on “objective” data, such as the next White House lie to break into the news. Others will be based on scenarios that may be revealed over time via investigative journalism, congressional hearings, and other techniques, such as how and at what stage the White House officially decided to use the attacks of September 11, 2001, as a reason for invading Iraq.

Two other markets are offered as entertainment and actually serve more analytic intentions. The experimental Foresight Exchange has traders betting on future events across a broader range of domains, including arts, business, government, and science. Long Bets is a spin-off of a nonprofit foundation whose mission is to foster better long-term thinking.

Behind corporate firewalls, businesses, too, are creating information markets. Hewlett-Packard is reported to have created one to forecast sales. A select number of employees, rarely more than two dozen, bought and sold shares depending on what they thought sales in a particular month might be. Each market ran for only a week. Over the course of three years, the information markets were more accurate more often than company forecasts prepared using traditional surveys and questionnaires.

Similarly, a pharmaceutical company is reported to have set up an internal information market to boost the efficiency of its research and development (R&D) spending. In this case, company employees, in addition to their conventional compensation packages, could invest in the progress of dozens of company drugs being developed by hundreds of employees, all competing for company resources. This front-line intelligence — aggregated, mutually assessed, and valued in stock prices — proved a more accurate monitor than management reports, software analysis, or committee meetings of which new drugs were making the most progress.

Other experimental markets enable consumers to trade the traits of products and services and allow companies to forecast more accurately the relative appeal of competing product feature sets. Trading does not measure and cannot replace measures of individual preference, so traditional methods will remain in demand. At the same time, these new information markets, which can stabilize, sometimes in less than 20 minutes, on the relative value of competing sets of product and service features, are likely to prove hugely valuable in both R&D and marketing.

It is no secret why these markets work. 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. Every market creates its value by aggregating little bits of information that don’t mean much on their own but become meaningful together. This reflects the network principle that the collective intelligence of a group is almost always smarter than the smartest individuals in them (for certain tasks).

In addition, the absence of hierarchy helps ensure no single person or viewpoint dominates and a diversity of viewpoints contributes to decision making. No person or viewpoint is shut out. Finally, information markets avoid the politics and careerism that so often block, divert, and channel the flow of information within organizations. Because people are rewarded only for being right, they have no incentive to hide information, pursue agendas, or go along with the crowd.

In general, any organization that can frame its informational needs as predictions can set up an investing game where the trading patterns of experts, front-line personnel, or customers will generate streams of highly condensed, real-time forecasting data that can be used to track the effect of current actions on future events. The Internet is an enabling technology for the Information Age, here as elsewhere. The more frequent use of this application is a sure thing. You can bet on it.