Manipulating perception
Intrade.com is a website where punters can buy ’shares’ on political outcomes. On the site, you can buy ’shares’ which gives you a dollar if say, Obama wins the presidency, and nothing otherwise. Before the event, the shares trade at prices between zero and one, giving the market’s perception of the probability that the outcome will actually occur. Justin Wolfers points out that the market was able to accurately call the 2004 Bush-Kerry Presidential elections, along with the Electoral College votes by state and also the 2006 Senate race.
An event that has been buzzing about economics related blogs is the news that an investor has been trying to manipulate John McCain’s share price. It hasn’t achieved very much - more rational investors quickly drove the price back down - but it surely Intrade’s coming of age if somebody finds it worthwhile to spend real money to manipulate the political betting market.
(Incidentally, the founder of Intrade.com lives in Ireland - online gambling is illegal in the US.)
any good summary of how the probability translates into the price? it seems to me that it follows if the market is risk-neutral (thus, market price = Prob * Face value) — but can be true only if there’s at least one risk-neutral investor with sufficiently large amounts of capital to back him up, and i don’t know who this guy is, or how we prove he exists.
Adrian
19 Oct 08 at 8:42 pm
Actually, the price remained out of line by about 8% for a period of around 3 weeks. There are a number of other markets that trade the presidential contracts like Betfair (doesn’t accept US bets), IEM and Hubdub (our exchange).
Even today there is some disparity in the price with McCain trading at 16% with Intrade and 12% elsewhere. Here is our market (with links to the other markets): http://www.hubdub.com/m17795/Who_will_win_the_2008_US_Presidential_Election
Nigel Eccles
20 Oct 08 at 1:07 am
Cheers Nigel. Hubdub looks like a lot of fun - I’m going to sign up for an account later.
I’m not doubting the accuracy of Hubdub vs other exchanges, but I’m wondering out loud whether using virtual or actual alters the accuracy of the predictions.
The intuitive answer is that the exchange with ‘real’ money is more accurate, but I’m guessing that exchange with virtual money is likely to get greater participation and people are more likely to be risk neutral (or even risk-loving).
Dom
20 Oct 08 at 5:03 am