Sep

27

 This site is devoted to the scientific method… expectations, the real world, actual decisions that people make under uncertainty. Any individual taking an economics course learns that consumer choice takes into consideration a myriad of expectations about the future subject to constraints and substitutions and alternatives. Please go back to the economics texts to see why prediction markets are much more accurate than polls. The prediction market is 75% for the incumbent. That's an all time high. Gentleman, does it have to go to 99% before you see that people actually making bets with their money is a much better predictor of outcomes than a poll? A good article assessing accuracy of expectations and margins of error for predictions versus polls is by Berg. Please. No more self supporting ideas about how close the polls are.

Stefan Jovanovich writes:

Ouch. Since all my ideas are self-supporting, I can only confess to absolute guilt. I also have to agree that money is and should be the litmus test. Intrade is not about money, however. The current "market" for Obama is 1 share offered at $7.51 and 104 shares asked at $7.46. Their comment stream, on the other hand, is unending; it dwarfs even the Huffington Post in frequency. Polls matter precisely because they are about money. They are the only device the campaigns (including the supposedly independent issue ones) can use to decide where to spend their advertising dollars and where to schedule the candidate appearances. Professor Berg's assertions about Intrade's markets are 10 years old; they also go back to the golden age when those markets themselves were so obscure that they were, indeed, pure. They are anything but that now, even if they remain shallower than the Platte River in September. I suggest that we all look at these questions the way the advertisers and producers look at audience ratings in television and radio; the overnights matter but P&G, Colgate and the car companies all want to see the internals so they can decide where to put down their next bets. What everyone knows is that trusting the raw numbers during sweeps week is not the best way to decide how to spend the hundred million dollars required to launch a new household cleaning product.

Jeff Watson writes: 

If polling offers more predictability than incentive markets, then perhaps one should look at the paper traders for guidance in the markets.

Rocky Humbert writes: 

Unless someone changed the law when I was not looking, it is unlawful for a US Citizen to bet on Intrade. When I tried to open an Intrade account several years ago, this fact was made very clear to me by the Intrade people. (And I didn't open an account.)

Hence you either have the US Election being predicted/decided by non-voters. Or you have the US Election being predicted/decided by Americans who flaunt the law.

I report. You decide.

Jason Ruspini adds: 

Liquidity used to come in during US hours and looking at just the past two days for the Obama contract, that still seems to be the case. The federal law that might be most relevant for listers is the Commodity Exchange Act. With Cantor movie futures and Nadex, the CFTC signalled jurisdiction over prediction markets, which would make Intrade an illegal commodity exchange. I guess they are busy with other things…

I have a theory that the hassle of wiring money in clips of less than $10k coupled with the margin system (you post $6 to buy a 60% contract but $4 to sell it) means that not only are the markets thin, but prices tend to be closer to 50% than they otherwise would, beyond the usual longshot issue near extremes.

EDT Hour    Volume

0                111

1                 36

2                 193

3                 60

4                 198

5                 283

6                 148

7                   22

8                  297

9                  537

10                270

11                3334

12                6621

13                1883

14                3079

15                2819

16                  262

17                8171

18                1961

19                6897

20                 101

21                 346

22                 536

23                 400
 


Comments

Name

Email

Website

Speak your mind

4 Comments so far

  1. steve on September 27, 2012 11:55 am

    With all due respect Chair you bring out a statement that this so called prediction poll is an an all time high. So has this so called prediction poll been tested against past polls. What is your statistical support to enforce such a dramatic statement on your part?

    Let me make some anecdotal comments. Last year who was the winner of the Super Bowl? Answer what was the early line of winning the Super Bowl before the season began before the playoffs began and who was the favorite for the Super Bowl. That is why they play the games.

    How many times does the Kentucky Derby favorite win the race. That is why they race the horses.

    Dewey Wins. OOps
    in 2000 Major networks publicly stated that based on exit polls, Gore wins Florida. The panhandle had yet to close their voting polls. OOps

    Bill Clinton comes out of nowhere in 1992 and wins the Primary and the General.

    Barack Obama wins the primary against the Clinton machine and by all accounts was declared dead. OOps

    Mark Twain “Rumors of my demise have been greatly exaggerated”

    Figures lie and liars figure.

    40 days to go. It is not over until it is over and one monkey does not close the show.

  2. Stefanie Harvey on September 27, 2012 12:38 pm

    There is an interesting website on UK elections, called Electoral Calculus.

    Here is a link discussing their 2010 track record using three main models (with two variants)

    http://www.electoralcalculus.co.uk/trackrecord_10models.html

    Note the complexity. How can this type of complexity be captured in a poll?

  3. vic on September 27, 2012 4:52 pm

    one important point about prediction markets is they take into account how the particiapants believe others will vote. a la keynes’ beauty contest in the newspaper where the idea is to predict who will get the most votes for beautiful women, not who you think is going to win.that’s what the stock market is about and what prediction markets capture. Keynes record at investments is apparently fairly good. he was always long, and he did not deduct his costs of entertaining “jocular ” friends from the pot. vic

  4. Terry on October 1, 2012 6:51 pm

    Without question, ‘money talks, and phone call polling walks.’ However, we also know that all of the polls are sampling the turnout of 2008, and that turnout will not be the same as 2012.

    This I know for a certainty. There is a anger and passion that middle America will show on election day, in an attempt to remove the current administration. We are not nuanced out here in the fly over states.

    Our America is in our hearts, it means something to us, and we show up and vote like never before.

    (from a small business owner in Indiana and the Netherlands)

Archives

Resources & Links

Search