Two days before the election, I call upon the media to dump the polls. Forget the discussions of the Bradley effect or the reverse Bradley effect and go to something that while not infallible, has proven much more reliable than polls. Look at the predictive markets!

Predictive Markets allow people to put real money on the outcome of the election. The two most popular are the Iowa Predictive Markets and Intrade.

I am going to review two markets from the Iowa Predictive Market over the past two Presidential elections. One is called vote share, where market participants attempt to nail what share of the vote the candidates will get. The other is called Winner Take All. this is simplest, where you put your money on who you think will win.

Two days before the 2000 election

Vote Share Prediction
Al Gore – 49.6%
George Bush – 49.4%

Winner Take All
Al Gore – 27.3%
George Bush – 74.9%

This market had it almost dead right. Al Gore got 48.4%, George Bush 47.9% but Bush won.

Two days before the 2004 election

Vote Share Prediction
George Bush – 51.7%
John Kerry – 48%

Winner Take All
George Bush – 55.2%
John Kerry – 45.3%

Again very close on the vote count which ended up with George Bush getting 50.7% and John Kerry pulling down 48.3%

So where does the current election sit in the predictive market two days before the 2008 election? As of this writing here are the figures.

Vote Share Prediction
John McCain – 47%
Barack Obama – 53.5%

Winner Take All
John McCain – 11.2%
Barack Obama – 88.3%

Remember that the markets only predict what is likely to happen if the vote were taken at that moment. So the day of the election is the day to watch. Of course, it’s possible that the predictive market could suffer it’s first major error. If not, it looks right now, as if Barack Obama will be the next President.