At some point in the past 20 years the interested amateur began to struggle to keep up with economic theory. It was just too hard to enjoy the latest academic work unless you had a head for higher mathematics. Recently, however, some writers, notably Paul Krugman in the New York Times, have been trying to drag economics back into the mainstream. The subject needs to stay there. As Thomas Sowell, a professor at Stanford University’s Hoover Institution, argues, economics lies at the heart of many political issues. – quoted from “The importance of economics: A black and white case”, The Economist, January 5th 2008

As someone who is in a somewhat frusturating pursuit of a Masters in Economics right now, this quote struck a chord. Not only does it seem you need a head for higher mathematics to keep up with the subject, those that study it can been forgiven for asking if that is all you need – sometimes it seems that class after class tests and rewards hopeful economists on little more than their ability to manipulate ever more complicated equations, in hypothetical situations that seem to get further and further from reality rather than closer to it as time goes on. While there are certain benefits that can come from this particular approach, there are also many costs - particularly in relation to all the bright people that flee from the field in terror and/or boredom, and the lack of development in reading, writing, communication, and interpretation skills for those that trudge through.

But that’s all part of a story for another time – at this point I bet you are asking what the heck it has to do with Google’s prediction markets. Well, online prediction markets are something that many economists absolutely love for a number of reasons  – notably, set up properly they can create an accurate picture of “market” expectations for any any number of future events (economists tend to love markets), in addition to perfect information on each person’s bets within the market, which can be cross-tabbed with all kinds of other information about the market participants, which let’s you test all kinds of crazy things and make many inferences about why and what people do.

At the same time, normal (I thought about putting this word in quotes, than decided against it people not only easily understand prediction markets, many love them – gotta dig the Kristanna Loke Starbond going up $2.66 today, non? But even if you don’t participate in the Hollywood Exchange, most can understand prediction markets with one word – gambling – and who doesn’t understand gambling?

In turn, studies on prediction markets can not only teach us all kinds of new things about human behavior and (say) how networks operate, they can help bridge the gap that’s divided economic research and the mainstream, and while I’m not directly emphasizing it in this post there are rather obvious business implications. This paper on a huge Google experiment with prediction markets is the biggest and best to come along yet on the topic. I was going to summarize it, but this NY Times article does a great job so it seemed like a waste.

Actually, I’ll list some of the more interesting findings anyway (jist of the experiment – Google folks bet on stuff that would happen in the company and the people running the experiment looked for patterns in the predictions in relation to other variables): 

• internal markets overpriced securities tied to optimistic outcomes by 10% – and this optimistic bias was significantly higher on days that followed Google’s stock going up
• securities tied to extreme outcomes were underpriced slightly (something the author of the Black Swan knows all about)
• corporate prediction markets appear to perform better as collective experience increases
• most measures of demographic similarities seemed to have no effect, but sharing a common non-English native language appears to count for something
• proximity matters – to a point. Those that share an office have highly correlated prediction patterns, while those on different floors might as well be in different cities.  

If you want to dig a bit deeper into this topic, the Wikipedia links and the studies that come up on the first few pages of a simple Google search for “prediction markets” will take you a long way, and Inkling is a great place/ company to look at in terms of how a business can harness the power of prediction markets - now I’ve got to get back to totally differentiating some funny looking equations.