Prediction Markets and the Wisdom of Crowds

Here's an article in the Economist that caught my eye (I recently subscribed to the Economist, does it show?) It caught my eye because it mentions Koch Industries, the founder of which funded the Charles G. Koch summer fellowship that I participated in last summer. Koch Industries is a fascinating company, and their CEO is a big fan of Austrian economics and free markets. One day I'll probably write a post on him and his company.

Anyway, true to form, Koch Industries has apparently been using prediction markets within their company for a few months to determine the course of various outcomes, from commodity prices to bailouts. Prediction markets, for those who don't know, are markets one sets up in which people buy shares in a particular outcome. As an outcome becomes more popular (people think it is more likely) the price goes up. Koch, as well as some other firms, have been using these internal prediction markets to forecast.

I find the idea of prediction markets really interesting. On the one hand, few of these people have any considerable knowledge of the subject in question. On the other hand, the law of large numbers suggests that if you pool enough people, you'll get something resembling the right answer. Koch says it has been working pretty well. I think something like commodity prices would require at least some area of expertise, but assuming that, I think running a prediction market can be a really effective way to create incentives for accurate forecasting.