Tuesday, October 2, 2012

Romney should buy shares in Obama

Over at Kunstler's blog, right after calling Willard the Rat "the biggest empty suit I've seen in 44 years of following politics," I noted that I had "stayed away from the presidential race this past week" here. Time to change that with a piece about my favorite political market, the Iowa Electronic Market (IEM), which has been in operation since 1988 and that I've been following since 1992, 20 years ago.

Venture Beat: Polls show a bounce – but the smart money has always been on Obama
Brent Stinski
September 18, 2012 11:00 AM
Ask people to bet on a candidate, to put down real cash on who will really win – and the smart money, for some time, has been on Obama. Certainly that’s the case at so-called prediction markets, forums where people can place bets, often with real currency, on what will happen in November. As of this writing, odds from three independently run online markets present a 65-plus percent probability that Obama will win, including odds at my own startup, Media Predict.

That’s no slam-dunk. Still, Obama has a clear lead. In fact, according to the markets, he’s had that advantage since the start of his campaign. But mainstream media only seem to be waking up to this fact in light of a recent post-convention “bounce” – endorsed by, yes, those beloved, slow, traditional polls.

Well, welcome to the party. For those who will want to remain a step ahead in this campaign, we recommend looking to the markets instead. Their message is clear and strong.

Media Predict’s 64% probability of an Obama victory is virtually identical to those furnished by other forums, such as the Iowa Electronic Markets (a real-money futures market the CFTC permits for purposes of academic research) as well as offshore gambling sites like Intrade. And there’s good reason to believe what the markets say. A 16-year retroactive study found the Iowa Electronic Markets to outperform the AP and the Gallup Polls 74% of the time when predicting presidential vote share on the eve of the election. Moreover, the study found the IEM to be almost always more accurate when predicting from 100 days away or more – a point we passed some time ago.
When I included this article in the tip jar to Overnight News Digest: Science Saturday (Streambed on Mars discovered edition) on Daily Kos, the current quotes at the IEM showed Obama with an 80.6% chance of victory and 53.7% vote share, which were almost identical to the probabilaties at Nate Silver's FiveThirtyEight blog. Since then, the two measures have diverged, particularly on the likelihood of election. The IEM now lists Obama with a 77.5% chance and a 53.9% vote share to Romney's 23.0% chance and 45% vote share. In contrast, FiveThirtyEight's forecast gives Obama a better chance of victory at 85.7% but a lower share of the vote, 51.5%. However, the shared pattern of the election forecast remains the same--a steady race until after Clint Eastwood yelled at a chair, then rapid divergence the month after that.


If I were Romney, I'd buy shares in Obama.

No comments:

Post a Comment