This is part one of a two-part series.
My sister-in-law gave my husband and me a subscription to Wired Magazine last holiday season. The gift was thoughtful. We both write a blog: he writes SpendMatters and, well, you are reading the one that I write. We both enjoy our eclectic friends from all over the world in various industries. Our kids inform of us of the latest shows; our friends show us the latest in internet gaming. So when I stumbled upon an article titled The Crowd Calls it Wrong, I became very intrigued. Of course, the article rubbed it in a little, “As you’ve no doubt heard, prediction markets are online trading sites that let people buy and sell shares tied to, among myriad other things, the fortunes of candidates and parties…”
I admit, I had heard snippets about these predictive markets. I remember reading about something Google did internally. Unfortunately, I didn’t really connect all of the dots. My tech-savvy husband, though, knows them fairly well, and mentioned that there are companies who license such applications for as little as a few thousand dollars a year. Or, one can try an open source model such as Zocalo.
What are predictive markets? According to the Wired article, “prediction markets are big information processors, distilling the collective wisdom of their traders.” The article goes on to discuss how the predictive markets, and one in particular, The Iowa Electronic Markets has proven itself more accurate than the political polls at least 75% of the time. But they have also made some big errors, in this case, they completely miss-called the Democratic New Hampshire primary (predicting a big Obama win when instead Hillary won).
Did you know that any and all of us could participate in a predictive market to say whether or not Eliot Spitzer will be indicted on felony charges in 2008? Or whether or not there will be a 9.0 magnitude earthquake before the end of the year? Or if mad cow disease will be be contracted here in the US before the end of 2009? These are just some of the many predictive markets available via Intrade, one of the bigger predictive market sites.
Mysteriously absent based on my own searching are any predictive markets discussing the future pricing of steel, copper or aluminum. The best I found related to an experiment conducted by steel giant, ArcelorMittal. According to this Newsweek article, ArcelorMittal tried to predict sales volumes and pricing of a key product in September 2005, in this case hot rolled steel. The result? They got accurate predictions on volume, but not on price. The Newsweek article goes on to say that predictive models may work better inside of companies … to give everyone a voice so to speak. And they may also work better for technology and consumer packaged goods companies.
We’ll cover what is necessary to make these markets work in a follow-up post.