Berg, Rietz Say IEM Could Help Identify Best Candidates
The IEM, a real money web-based market operated by six professors at the Henry B. Tippie College of Business at the University of Iowa, has a reputation for accurately predicting elections since its inception in 1988, with an average margin of error of 1.37 percent.
IEM co-directors Joyce Berg, professor of accounting at the UI, and Tom Rietz, associate professor of finance, say the IEM can be used by political parties to make decisions based on the prospects of their candidates. In particular, political pundits can look at IEM's presidential vote share market, which stacks up Democratic candidates against George Bush.
"At this point, the IEM shows that Dean is not a strong candidate to run against Bush," Berg said. "There are other Democratic candidates that are predicted to fare better against Bush."
Traders in the IEM use real money to back up their choices, unlike opinion polls, where respondents have few incentives to tell the truth, Rietz said. Values of traded contracts depend directly on future outcomes, so prices give information about these outcomes.
In their paper published earlier this year, "Prediction Markets as Decision Support Systems" Berg and Rietz give an example of how the IEM could be used for supporting political campaign decisions. According to Berg and Rietz, the IEM vote-share markets for the 1996 Presidential election showed that Colin Powell would have been a particularly strong candidate against Clinton, while Robert Dole was a particularly weak candidate against Clinton.
According to IEM market data, Clinton's chances of winning the election increased with the probability that Dole would be the Republican nominee. The opposite was true for Powell: Clinton's chances of winning dropped as Powell's chance of being the Republican nominee rose
"When Powell dropped out of the race, Clinton's chances of winning jumped in our market," Berg said.
"Had the Republican Party or even large numbers of primary voters conditioned their decisions on market information, they could well have nominated a stronger candidate, " Rietz added.
A complete copy of "Prediction Markets as Decision Support Systems" is available online.
Contact: Joyce Berg, Iowa Electronic Markets, 319-335-0840