Sunday, September 20, 2015

Why Prediction Markets still have a future

"Why Prediction Markets still have a future"
By: John Authers
Financial Times  -  March 11th, 2013



There are numerous prediction markets today, however in 2013 the best prediction market according to this writer for the Financial Times is "Intrade". Unfortunately it closed in 2013 and the author wanted to examine whether prediction markets would continue to grow or whether it would decrease in popularity.

First he explains what prediction markets are. Quite simply, it is a system that allows people to stake bets, both real or imaginary, in predicting an outcome. Th example he gives is sports betting, which allows gamblers to stake bets on both the outcome, point spread, and other factors related to sports. It is also used in the financial world in predicting outcomes. At the time the Wall Street Journal even published the odds related to trading so investors could hedge their bets.

As expected, a common usage of prediction markets is predicting political results. The interesting tidbit of this is that political prediction markets date back to the turn of the century. In the early years of the 20th century, it is estimated that prediction market gambling on political results generated $37 million in bets. It should also be noted that these political prediction markets all outperformed the Gallup poll in predicting the political outcomes. The author does note that while these prediction markets are usually very accurate, they are only as reliable as the information that is available and fed. For example, Intrade wrongly predicted that the US would discover WMDs in Iraq, but as we know in hindsight that was due to faulty information. To put it simply, the author believes prediction markets are too useful to go away.


I think the author could have done a better job in explaining exactly what type of information is fed into a prediction market. The author in my opinion explains it scantly and makes prediction markets appear to be random betting like amateur gamblers at a horse race. A random gambler may take a chance on some stock that he or she feels will be profitable, a true prediction market expert will have reliable information that exceedingly increases the odds in their favor. I do find it interesting that prediction markets are more accurate than the Gallup Poll in politics, but that makes more sense in that the information fed into prediction markets is far more nuanced than random opinions by people, especially in light of our complicated political system. I do think in hindsight he is wrong partly, sports betting has taken off however I did not have any knowledge prior to this class about prediction markets, and in politics the Gallup Poll is still widely used and respected.


  1. Since prediction markets rely heavily on the information fed to them, would this also increase the risk of biases occurring?

    1. Katie, while biases are definitely a concern, I think the way the market functions, through changing prices that reflect investor/analyst confidence, tends to average out, or at least reduce, individual biases. My main concern, is that the more difficult questions we have to answer typically have false or incomplete information, such as the Iraq WMD question. Can this method truly be called effective if it fails with insufficient information?

    2. Yes, although I think some industries are less prone than others. The financial industry for example I believe would be less prone due to the lack of emotion, whereas something like sports can easily be more emotional and lacking in pure raw data.

      And as Andrew said, it's not perfect but it's pretty good. If most of the known evidence pointed to no WMDs, it would not have predicted it. It can't overcome large-scale human error, no analytical technique can.