Andrew Leigh and Justin Wolfers have made a useful contribution to the ongoing debate: what is more effective at forecasting elections, betting markets, opinion polls, or econometric models based on the state of the economy? Looking at the the 2004 Australian federal election, their new paper Competing approaches to forecasting elections concludes:
While the evidence for economic voting has historically been weak for Australia, the 2004 election suggests an increasingly important role for these models. The performance of polls was quite uneven, and predictions both across pollsters, and through time, vary too much to be particularly useful. Betting markets provide an interesting contrast, and a slew of data from various betting agencies suggests a more reasonable degree of volatility, and useful forecasting performance both throughout the election cycle and across individual electorates.
John Quiggin comments that polls and betting markets "are about equally good, at least when an election is about to happen." He continues:
There’s no indication that markets have significant private information: for example, they react, and sometimes overreact to ‘news’ that turns out, in retrospect, to be misleading. But most of the time, they provide a pretty good summary of available public information.
Maybe so, but one would have thought key party insiders have access to better political poll data than the rather small national surveys published by newspapers: clear potential for insider trading. Perhaps the data is too volatile, or swinging voters too numerous, to resolve the 'noise' problem?
I've watched the tradesports election data very carefully for a while now. For the 2004 presidential election I watched the data daily for a year, hourly for the week before the election, and minute by minute in the 48 hours leading up to and including the "call" of the election by the relevant state level election officials.
Here's what I learned:
1. In the '04 election, those entering the market early (like a year before the election) tended to be irrationally exuberant Republicans. For example, the contract for BUSH to win Oregon was priced at 60 (or so) a year before the election -- and of course Bush lost Oregon and the contract closed at zero. Every state with a seeming mismatch between polling data and trading price was in the same direction -- irrationally priced for Bush to win.
2. As trade volume increased, the prices became more rational. (I.e. the price for Bush to win Oregon approached, and fell below, Bush's polling numbers).
3. On the eve of the election, the data swung wildly as traders responded to scraps of news. The big swings came when exit polling data was leaked and spread on the internet -- suddenly Kerry WOULD win Ohio! Ka-ching! And then the prices dropped as the polling was disputed and disproved by early election returns in key precincts.
I'm sure Wolfers et. al have taken a systematic look at this data. It would be interesting to see what they can tease out and illuminate with the data, and what is obscured by the choice of model.
Posted by: decon | Wednesday, March 30, 2005 at 11:14 PM
There's clearly some insider trading going on, based on party-polling. I know, I have friends who did it (and the party bosses seem don't really seem to care). My suspicion is that many of those placing seat-by-seat bets are either party insiders watching the internal polling, or activists working on those campaigns - which is why those markets tend to do better than experts. Similarly, in the last US election, my understanding is that the Iowa favourite won in all 50 states. This kind of local data-aggregation actually seems to me to be one of the best aspects of betting markets.
Posted by: Andrew Leigh | Tuesday, April 12, 2005 at 12:29 AM
Hello Great post about Forecasting elections - polls vs markets. Really a very nice piece of information. Newbies like me get a good idea about this... I'm going to comment now... I think i did it. Thank you..
Posted by: Buy Viagra | Wednesday, December 09, 2009 at 04:52 PM
istanbul hotel molebe kansas artic thanks This article is very beautiful, I really get very beyendım text files manually to your health as you travesti very beautiful and I wish you continued success with all respect ..
En güzel travestiler
travesti çok yakında burda yayımlanacaktır bizi izlemeye devam edin
Thanks for helpful information travesti siteleri you catch up us with your sagol instructional çok explanation.
en iyi travestiler en guzel travesti
travesti forum
istanbul travestileri
ankara travestileri
izmir travestileri
bursa travestileri
travestiler
trv
travesti siteleri
travesti video
travesti
travesti
travestiler
travestiler
sohbet
travesti
chat
organik
güncel blog
sohbet
turkce mirc
çet
Posted by: travesti | Saturday, July 03, 2010 at 11:50 AM
Thank you for this wonderful article ... really very nice - there are such things
amatör şairler
Esrarlı sözler
güzel sözler
Gooo
isimler
Sohbet
Posted by: aşk sözleri | Friday, August 13, 2010 at 08:56 AM
Çilek sex shop Mağzamızda En Kaliteli sex ürünleri, sex oyuncakları şişme bebek ve erotic shop erotik giyim Ürünlerini Bulabileceğiniz Gibi Ayrıca, penis büyütücü, geciktirici, Bayan Uyarıcı, Ürünlerde Temin Edebilirsiniz. 1994 ten Bu Yana En Kaliteli Orjinal Erotik Ürünlerini Sağlamakta Olan Çilek Erotik Shop ta Tüm Cinsel Ürünleri Bulabilirsiniz
Posted by: Maria | Thursday, February 17, 2011 at 01:09 PM