Friday, September 28, 2012
Prediction markets and elections, 2000-2012
I'm not a campaign analyst, but I have dabbled in prediction markets, which means I've kept track of the Iowa Electronic Market.
If we look at the current election, we see that over the past month the aggregate prediction of the probability of Obama winning broke through the bounds that it had operated within since last July.
This looks pretty dramatic, but let's look to history. In 2008 the Obama contract had a breakout past the 70% mark around October 4th.
In 2004, the Bush contract peaked somewhere in the low 70% range around the end of September.
In 2000, the Gore contract had a local peak in late September.
This rough historical comparison indicates that now is about the point in the 2008 election when the market began to show Obama pulling away. In 2004 we saw a reversal of a sharp upward trend in the leading candidate, though the leader pulled out the victory in the end and the reversal had already begun at this point in the election cycle. In the wild 2000 election, we saw Bush open up a lead in mid-October, but that lead never broke the 80% mark.
If we look at the current election, we see that over the past month the aggregate prediction of the probability of Obama winning broke through the bounds that it had operated within since last July.
This looks pretty dramatic, but let's look to history. In 2008 the Obama contract had a breakout past the 70% mark around October 4th.
In 2004, the Bush contract peaked somewhere in the low 70% range around the end of September.
In 2000, the Gore contract had a local peak in late September.
This rough historical comparison indicates that now is about the point in the 2008 election when the market began to show Obama pulling away. In 2004 we saw a reversal of a sharp upward trend in the leading candidate, though the leader pulled out the victory in the end and the reversal had already begun at this point in the election cycle. In the wild 2000 election, we saw Bush open up a lead in mid-October, but that lead never broke the 80% mark.
# posted by Wiggins @ 10:35 AM 
