What are the Iowa Electronic Markets, and did they really beat the polls?
Since the 1988 election, a university has run a tiny real-money market where you can risk at most $500. It topped the polls 74 percent of the time in one study.
Guess first
?
Which one is it?
The Iowa Electronic Markets, or IEM, are real-money prediction markets run by the University of Iowa's business school. They were introduced during the 1988 US presidential election, and they are not a casino: the markets are and run for educational and research purposes. In one large study they beat the polls about three times out of four, though that is not the whole story.
The idea began as an academic experiment. A , the Iowa Political Stock Market, was invented by George Neumann and developed with Robert Forsythe and Forrest Nelson. Traders buy and sell contracts tied to things like election results and , and the price of a contract is read as the chance of the outcome.
The stakes are deliberately tiny. A trader can put at risk only between $5 and $500. Because of that academic focus and those small sums, the IEM is not by the US Commodity Futures Trading Commission or any other agency.
Did it really beat the polls? A study compared 964 election polls with the five US presidential elections from 1988 to 2004. Berg and colleagues showed that the IEM topped the polls 74 percent of the time. In 2008 it predicted the final vote count to within half a percentage point. But the researchers Erikson and Wlezien (2008) compared the IEM with "damped" polls, poll numbers adjusted before being read as a forecast, and found the damped polls outperformed all markets or models. Other work found that polls weighted by how confident people were performed as well as market forecasts, so markets are one good tool, not a magic one.
Quiz me
0/3
Recap
In one study of five US elections the IEM topped the polls 74 percent of the time, but Erikson and Wlezien found adjusted, damped polls did better still.
💡 A trick to remember it · A small pot, a big question: $500 at most, and still it often beat the pollsters.
Surprising fact · A trader can risk only between $5 and $500, which is why it is not regulated like a normal exchange.
Connects to
- 🗳️ Are prediction markets more accurate than polls?
- 🐂 Can a crowd of ordinary people guess better than the experts?
- 🎯 What does it mean for a forecast to be well calibrated, and how is it scored?
- 🔮 What is a prediction market, and how does a price become a probability?
- 🗳️ How can a poll of a thousand people speak for millions, when one of 2.4 million got it wrong?
- 🧭 Why do smart people make the same mistakes in the same way?
Sources (3)
No source, no claim. Every fact in this lesson (12 claims) cites at least one of these.