« Tying Odysseus to the mast | Main | ICD: A 5-step program? »

February 04, 2008

The Industry Standard returns as prediction market

The Industry Standard was one of the go-to news sources during the late 1990s dot-com boom. Like so many of the companies on which it reported, it went bankrupt in 2001. However, the computer industry trade publisher International Data Group brought it back to life today.

In a novel twist, The Industry Standard is now combined online newspaper and a prediction market. With a free account, readers can place (fake currency) side bets on various predictions about industry events. For example, markets open today include whether Yahoo! will accept Microsoft's bid by the end of the week, and whether Google, Yahoo! or Microsoft will buy Tivo by the end of the summer.

Prediction markets are a type of incentive-centered design that have become widespread and hot in the past few years. The first significant example I recall is the 1988 Idea Futures market by Robin Hanson, though academic articles by Hofstee (1984) and Leamer (1986) came a bit earlier.

There are quite a few well-known and active prediction markets now, such as Hollywood Stock Exchange, TradeSports, and News Futures. However, the new Industry Standard is the first time I've seen a significant market that is so closely linked to an online news source (though others sometimes provide a limited news feed). It seems like a natural idea: an industry-focused news site (if successful) is bringing in knowledgeable people with an interest in the likelihood of industry events.

Why is a prediction market an incentive-centered design? It is providing incentives to induce knowledgeable people to make the effort to reveal their knowledge and beliefs, and to do so honestly (so they address problems of both hidden action and hidden characteristics). Traditional stock markets provide this information revelation and aggregation role for the futures of publicly-traded companies; prediction markets play the same role for other outcomes.

Most of them, due to gambling and securities laws in the US, run based on funny money (valueless currency). Thus, the substance of the incentives is not immediately clear, and it is a topic for future research to determine how well these markets do induce effort and truth revelation. (The Iowa Electronic Markets, which has approval to trade in small amounts of money, mostly on political outcomes, has performed extremely well -- almost always better than professional polling organizations such as Gallup -- for many years.) It is going to be interesting and instructive to analyze participant responses to non-pecuniary incentives.

Former Michigan Ph.D. student Dave Pennock (now with Yahoo! Research) publishes an excellent blog about predicion markets: Oddhead.

(Image courtesy of "Financial Aid Podcast" on Flickr.

Posted by jmm at February 4, 2008 08:28 AM

Comments

Login to leave a comment. Create a new account.