Reporters, bloggers and political junkies love to cite prices on Intrade as indications of the true state of the Republican primaries. It’s understandable why–Intrade often gives a much clearer picture of each candidate’s chances than polling or political commentary. But Intrade has a number of flaws that aren’t apparent to passive observers. I started trading on Intrade last fall and was pretty surprised at the problems in what’s generally thought of as the top political predictions market. In short, the problems are illiquidity (leading to unreliable and volatile “last prediction” quotes, even in popular markets) and uncertainty regarding how Intrade is run and how it functions:
Spreads on Intrade are huge: For nearly all markets, there’s a significant spread between the standing buy price and the standing sell price. For example, if you want to go long Newt winning South Carolina, a contract will currently cost you $1.87 (18.7%). If you want to short the same contract, you’ll currently have to short it at $1.30 (13%)–a massive 50 cent spread for what should be one of the most popular contracts. While reporters and bloggers only cite a single percentage reflecting the most recent transaction (“Romney at 70% on Intrade”), the bid-ask spread gives you a much clearer picture of the market.
Intrade’s management is unpredictable: Last November, I bet against Herman Cain dropping out of the campaign before the end of 2011. This obviously didn’t work out too well for me, but what’s interesting is how Intrade reacted to Cain’s announcement that he was “suspending” his campaign. Cain’s announcement came on a Saturday and throughout the whole weekend, people bought and sold contracts (I think the going rate was around 90%) while they waited for Intrade to decide whether suspending a campaign is the same as dropping out. All of the trading was purely based on uncertainty regarding Intrade’s judgement. On Monday, Intrade not only decided that Cain’s suspension was equivalent to dropping out but they also reversed all trades which occurred since Cain’s announcement. That Cain had dropped out was a no-brainer but that Intrade would reverse trades which had occurred between willing counterparties was pretty shocking to me. Not the best way to run a market.
Intrade’s website can get overloaded: During the night of the Iowa caucuses, the price of contracts for Santorum, Romney and Paul fluctuated wildly as breaking news favored one candidate or the other. At one point, Nate Silver pointed out that Ron Paul had hit 50% on Intrade–for Ron Paul bears, a great opportunity for a short. Unfortunately for those bears, Intrade’s website was massively overloaded that night and, at least for me, too slow to get any trades in. So on a primary night, just when you might want to turn to Intrade to get an idea of what’s actually going might be when the site’s at its most unreliable.
All of this isn’t to say that Intrade’s predictions should be ignored. They’re often one of the most accurate predictors available. But when you see an Intrade probability on some political blog, just know it comes from a highly illiquid, at times dysfunctional market and treat it with appropriate caution.