Comments on: Forecasting Presidential elections http://vivekmohta.com/blog/2007/05/10/forecasting-presidential-elections/ Vivek Mohta's blog Sun, 06 Jul 2008 02:48:14 +0000 http://wordpress.org/?v=2.5 by: Rob Chhabra http://vivekmohta.com/blog/2007/05/10/forecasting-presidential-elections/#comment-678 Tue, 09 Oct 2007 17:20:30 +0000 http://vivekmohta.com/blog/2007/05/10/forecasting-presidential-elections/#comment-678 I have read and seen data that show a market based approach to forecasting elections is most accurate. In particular a website such as tradesports.com has a had more accuracy in predicting the outcomes in elections than national polls. I have read and seen data that show a market based approach to forecasting elections is most accurate. In particular a website such as tradesports.com has a had more accuracy in predicting the outcomes in elections than national polls.

]]> by: Ben http://vivekmohta.com/blog/2007/05/10/forecasting-presidential-elections/#comment-413 Tue, 21 Aug 2007 19:32:29 +0000 http://vivekmohta.com/blog/2007/05/10/forecasting-presidential-elections/#comment-413 The reason in favor of prediction markets is that they should incorporate all the other information out there. Maybe polls are important or fundamentals, but at any given point, the participants bidders-sellers in the marekt should use those data to set the price. You really just need a few informed participants to get the correct price. The reason in favor of prediction markets is that they should incorporate all the other information out there. Maybe polls are important or fundamentals, but at any given point, the participants bidders-sellers in the marekt should use those data to set the price. You really just need a few informed participants to get the correct price.

]]> by: Vivek Mohta http://vivekmohta.com/blog/2007/05/10/forecasting-presidential-elections/#comment-399 Thu, 16 Aug 2007 06:21:20 +0000 http://vivekmohta.com/blog/2007/05/10/forecasting-presidential-elections/#comment-399 Ben, My impression is that prediction markets are effective in certain contexts and less so in others. In particular, I think prediction markets are better at integrating sentiment than polls a week before a general election. However, for a primary election months away, prediction markets are quite volatile, just like the polls. In addition, prediction markets require input from other people while the methods referred to in this post attempt to make predictions based only on the more stable "fundamentals." In the end, you might be right that none of the methods based solely on the fundamentals are compelling. Ben,
My impression is that prediction markets are effective in certain contexts and less so in others. In particular, I think prediction markets are better at integrating sentiment than polls a week before a general election. However, for a primary election months away, prediction markets are quite volatile, just like the polls. In addition, prediction markets require input from other people while the methods referred to in this post attempt to make predictions based only on the more stable “fundamentals.”

In the end, you might be right that none of the methods based solely on the fundamentals are compelling.

]]> by: Ben http://vivekmohta.com/blog/2007/05/10/forecasting-presidential-elections/#comment-273 Tue, 10 Jul 2007 19:34:55 +0000 http://vivekmohta.com/blog/2007/05/10/forecasting-presidential-elections/#comment-273 Read a lot of these papers in grad school. None were terribly compelling. If I had to predict, for the moment, I'd look at prediction markets such as the Iowa one. Wolfers has a paper on them. Read a lot of these papers in grad school. None were terribly compelling. If I had to predict, for the moment, I’d look at prediction markets such as the Iowa one. Wolfers has a paper on them.

]]>