April 12, 2010
IEM offers lessons in box office prediction markets
A division of the bond trading firm Cantor Fitzgerald is hoping to open its new box office prediction market this month, but the coordinator of the Iowa Electronic Market’s movie prediction market thinks traders will have their work cut out for them making money.
“Forecasting movies is a very difficult problem,” said Tom Gruca, a professor of marketing in the Tippie College of Business who studies the use of prediction markets as a forecast tool. “There’s an enormous amount of data out there for traders to aggregate, but how do you put it all together? Nobody’s learned yet how to distill all that information into a consistently accurate forecast.”
Cantor needs only one final green light from the Commodities Futures Trading Commission to move forward with its prediction market, although opposition from the film industry has currently delayed the project. Once approved, traders on the market will be able to buy and sell contracts based on the box office gross of a movie’s first four weeks.
The IEM, operated by the University of Iowa’s Tippie College of Business, has been operating a similar real-money box office market since 1995. The market operates primarily as a classroom market for one of Gruca’s fall semester graduate level marketing courses, so traders are made up mostly of students and other academic traders interested in learning about prediction markets. Traders buy and sell real money contracts based on their prediction of the opening four-week box office of a film Gruca selects.
The first market opened in 1995, shortly before the opening of another box office futures market, the Hollywood Stock Exchange, a for-fun only market that’s also operated by Cantor Fitzgerald.
Since then, Gruca said the two markets have shown how difficult it is to accurately forecast box office returns. The IEM’s average percentage error is 38 percent, while the Hollywood Box Office predictions for the same movies miss by an average of 31 percent.
Compare this to the IEM’s presidential vote share market, which has an average error of about 1.8 percent, and the difficulty of forecasting movie box office results is clear.
Like all IEM markets, traders on the movie market are limited to an investment of $500. Cantor’s market will have no limit, but Gruca said the track records of the IEM and Hollywood Stock Exchange suggest that cash doesn’t make much difference in a movie market’s predictive ability.
“The real money IEM markets do not improve on the forecasts from the Hollywood Stock Exchange, so financial incentives don’t play a big role in obtaining more accurate predictions,” Gruca said. “Neither of the movie markets does as good a job forecasting as the political prediction markets.”
Gruca wrote about the challenges of a box office prediction market in his paper, “Incentive and Accuracy Issues in Movie Prediction Markets,” published in 2008 in the Journal of Prediction Markets. The article was co-authored by Joyce Berg, professor of accounting in the Tippie College of Business, and Michael Cipriano, assistant professor of accounting at the College of Charleston in Charleston, SC.
STORY SOURCE: University of Iowa News Service, 300 Plaza Centre One, Iowa City, Iowa 52242-2500