February 2014

Sun Mon Tue Wed Thu Fri Sat


2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28  

Editor Login

Convener in chief:

David Lazer
(Methodology, Networked Governance)


Stanley Wasserman
(Current Trends, Methodology, Social Networks)

David Gibson
(Social Networks, Interaction, Theory)

Yu-Ru Lin
(Networks, Visualization)

Ines Mergel
(Knowledge Sharing, Social Computing, Social Software, Government 20)

Maria Binz-Scharf
(Qualitative Methodology, Knowledge Sharing, eGovernment)

Alexander Schellong
(Admin, eGovernment, Government 20, Citizen Relationship Management)



Recent Entries

Recent Comments


Powered by
Movable Type 4.24-en

View Blog Stats

Blog Directory & Search engine
Academics Blog Top Sites

Blog Flux Local - Massachusetts
Blog Flux Directory

« Social Network Feedback in Real Time | Main | Sunbelt 2008: Day 1 »

19 January 2008

Information, networks, and markets

It is notable, if not surprising, that the best place, in all of the Internet, to go for a sense of the direction that an election is going is intrade.com. Intrade is a futures market, where you can buy and sell contracts that pay off depending on the outcome of some event in the future. So, what has happened in the primaries is that that the futures market shows substantial shifts through the day of the event. Today is typical, where hours before poll closings, the market for the Republican race South Carolina swung hard toward McCain (who apparently is going to win).

What is interesting is that, as far as I can see, these markets swing before any other information is available on the Internet. That is, the market is being driven by those who presumably have a close connection to inside information—i.e. exit polls. This is not to say that employees of CNN are taking time off during the day to put money into intrade contracts, but that there is an informal leakage of information from those in the know, where that leakage then drives the market. That information can be misleading, of course (cf. Kerry, 2004); but the market allows for that possibility (e.g., even now Huckabee is trading at 2%).

This is an interesting natural experiment of sorts, because there those exit polls are likely by far the most reliable sources of information available regarding election outcomes (especially for primaries) before actual results come in. There are other bits of information that may be independently valuable, like any data on differential turnout across regions, etc, but generally the exit poll data are going to dominate everything else. It would therefore be interesting to calibrate the actual data that exit poll entities have gathered at what points in time (minute by minute) to see how quickly intrade responds. This should provide some sense of the speed of informational leakage from the exit polls. (And if you know of anyone has already done this, please post a comment.)

Btw, this graph shows the shift in the value of the McCain winning SC contract:


Note that the time marked is GMT, so the major shift occurred at around 6pm EST (exit polls were not released publicly until after 8pm).

Posted by David Lazer at January 19, 2008 9:12 PM