At this time of the year, lots of us start taking a closer look at the Intrade.com betting numbers. This is the site where people bet on who they think will win a certain question.
The current numbers for getting the nomination are 45.3% for Romney; 32.2% for Newt; and 6.2% for Huntsman. Ron Paul is at 5.1% and everyone else drops off from there.
My own impression is that these Intrade numbers are heavily influenced by not particularly critical conventional wisdom (not unlike financial markets); and they also tend in many cases to be lightly traded — thinning out some of the crowd-wisdom factor.
One other factor I’ve seen over the years is one political scientists will probably recognize in the internals of lots of partisan public opinion data: Republicans tend to be much more confident of political victory relative to what counts as relatively objective data and the historical record. Republicans just seem more temperamentally confident (and Dems the opposite) regardless of the political climate. That for instance made the chances of Democratic victory on Intrade seems far lower than I think actual public opinion data suggested. I’d be curious to hear from folks who work with opinion data whether they’ve seen good systematic data backing up this observation, or whether the observation seems valid at all.
Late Update: TPM Reader GB follows up with his experience with Intrade. His take matches broadly with mine.
! traded rather frequently on intrade in presidential nomination and general election markets in 2008.
My observations:
1. Intrade tended to be more reactive than prospective. New polls would come out, and intrade would react to them, rather than foreshadow them. 2. Intrade tends to like candidates who appeal to Wall Street or traders, and/or who have international name recognition (the company itself is based in Ireland). Rudy Giuliani was very slow to drop in 2008 even after it became quite clear he wasn’t going to be the GOP nominee.Jon Huntsman has typically been much better liked on intrade than in popular polls.
3. You’re right that the markets are rather small and lightly traded. The volumes they show are life-of-contract, not daily. But the market is small enough that a political consultant might consider tossing money in to try to move things, especially if more reporters keep citing the numbers. At current prices, by putting just $2500 at risk now, you could boost the bid price on Romney becoming the GOP nominee from 45.7 to 50 (although if/as more sellers came to market, you’d need to put in more money to keep it there).
4. An internal investigation at intrade from 2008 found that they believed someone was manipulating the market, trying to keep John McCain’s chances of winning the general election higher than the market itself would want:
So the idea that outside money might try to move the numbers isn’t just idle speculation, it’s happened before. And that was, in part, responsible for my trading success back in 2008.5. Finally, note that intrade is showing an expected probability of the person winning the election, which is different from showing support for a candidate in public polls. So these numbers should be quite different than polling numbers, especially in multi-candidate fields, or as a two candidate spread widens.
So yes, it’s real money at stake, which helps keep it grounded, but the sums are small enough that the market can be, and indeed has been, manipulated for a while if you’re willing to bear some loss. And the numbers to do that are small compared to budgets for a presidential election campaign.
Regards,
TPM Reader GB
