Recent Blog Posts
-
The Year in Research
Dec 31 20089:13 am EDT -
Mind Your Value Judgements
Dec 19 20087:52 pm EDT -
S.E.C. Short-Sale Ban: Pretty Much Useless
Dec 19 20083:45 pm EDT -
Advice from Japan: Don't Forget TARP 1
Dec 19 20082:31 pm EDT -
Chart of the Day: Money Market Stress Easing
Dec 18 20088:57 pm EDT
Links
- Junk Charts
- Economic Principals
- New York Federal Reserve Research
- Sabernomics
- Statistical Modeling, Causal Inference, and Social Science
- Sabermetric Research
- St. Louis Fed Research
- Bluematter
- NBER Working Papers
- TierneyLab
- Numbers Guy
- Social Science Statistics Blog
- DataPoints: The Dismal Scientist Blog
- Institute for the Study of Labor
- Predictably/Irrational
- Decision Science News
- Research Recap
- Econbrowser
- Center for Economic Policy Research
- Economist's View
- B.I.S. Working Papers
- Geary Behaviour Centre
- Real Time Economics
- Federal Reserve Working Papers
- C.B.O. Director's Blog
- Curious Capitalist
- VoxEU
- Freakonomics
- Philadelphia Fed Research
- O.E.C.D. Factblog
- MoneyScience
- Journal of Interest
- STATS Blog
- Email me
- EconTalk
- EconPapers
- Marginal Revolution
- Tim Harford
- Jeff Frankel
- Institute for the Study of Labor
- Social Science Research Network
Markets 2, Polls ... 2?
Despite the thin trading volumes, the prediction markets earned their name tonight by correctly picking Obama and Huckabee to win in Iowa while the polls showed much tighter races for both parties. (Although this last poll showed Obama and Huckabee leading.)
Looking ahead to next Tuesday's New Hampshire primary, the latest polls -- before the Iowa results -- show Clinton narrowly ahead. But the InTrade contracts have changed dramatically on the Democratic side tonight with Obama given a 53 percent chance of prevailing while Clinton's contract implies only about a 30 percent chance of victory. (Again the volume here leaves something to be desired.)
The much more liquid presidential nominee contracts on Intrade are also moving. Clinton's contract fell to 53 and Obama's climbed up to 43, its highest level.
On the Republican side, John McCain continues his surge and is now given a 65 percent chance of winning in New Hampshire.
And a correction to my earlier post on the markets and Iowa. I said:
Intrade has Obama by a large margin (60 percent to 16 percent for Clinton shortly past noon)
But that's not quite accurate. What the price of a contract in a prediction market tells us is the percentage chance that that event will happen. So the 60 figure you see whe you log on to InTrade means the market gives him a 60 percent chance of winning. Even if another candidate is only given a 10 percent chance of winning that doesn't mean that in the end the margin of victory will be as big as the differences in probabilities.
And who do I have to thank for setting me straight? Prediction market pioneer Justin Wolfers, who better explains:
A 60% chance v. a 16% chance may or may not translate into a large margin. For instance, I'm 100% certain that my brother is faster than me over 100 yards. But there's a chance that I'll be really quick today, so perhaps he's a 90% chance to beat me in any race. Equally, he's not much faster - the margin is typically about one foot. So a large likelihood here can translate into a small margin.
For what's sure to be the best analysis of the markets and the 2008 elections, I'd recommend reading Wolfers at NYT here and WSJ here. All I can ask is: What's next? USA Today?
