Prediction Markets: Research and Limits

Fact: A story in Science Daily this week, “Election Forecasters Preparing For Historic Election,” relates the publication this month of the “assembled insights of prominent election forecasters in a special issue of the International Journal of Forecasting.” 

Analysis:  The journal articles are available here for download.  One of them, “Prediction Market Accuracy in the Long Run” (by Joyce E. Berg, Forrest D. Nelson, and Thomas A. Reitz from the University of Iowa’s Tippie College of Business), compares the presidential election forecasts produced from the granddaddy of them all, the Iowa Electronic Market (IEM), to “forecasts from an exhaustive body of opinion polls.”  Science Daily says they find that the IEM is “usually more accurate than the polls.”

If we extrapolate out, these election markets are special cases of prediction markets, and I’m always interested in those.

A commenter on my post about Microsoft Research’s predictive algorithms has tipped me to a solid blog on prediction markets: MidasOracle.org, which is maintained by Chris Masse, whose personal site is a helpful set of definitions and descriptions of varying schools of thought and prediction approaches.  Masse, an American who lives in France, has posted recently a good list of different labs doing prediction-market research (including Microsoft Research).  He also maintains an updated list of various software products and packages (some are SAAS) in the field.

Controversial Incentives: Some kinds of prediction markets may create controversial incentives. For example, a market predicting the death of a world leader might be quite useful for those whose activities are strongly related to this leader’s policies, but it also might turn into an assassination market.”  -From the Wikipedia “Prediction Market” page

If you would like a basic introduction to the field, though, try checking out first the Wikipedia entry on prediction markets, which is replete with discussion of the controversies and sub-optimal aspects (both theoretical and practical) inherent in many prediction markets.  It is obviously a field with many holders of strong and conflicting viewpoints.

I guess that’s what makes a horse race…  But for the 2008 elections: while reading the Science Daily story, I predicted correctly that an academic would say something like this: “Forecasting this election will be more difficult than usual,” according to Prof. James E. Campbell, chair of the Department of Political Science at the University of Buffalo, “who since 1992 has produced a trial-heat-and-economy forecast of the U.S. presidential election.” 

And of the six different forecast models described in the journal articles, “only two have a forecast at this point. The other four will have forecasts between late July and Labor Day.”  Alert the media!

 Email this post to a friend