Predictive Markets: Capitalism gives better information

In a recent article, Moneysense magazine discusses predictive markets. These are information-gathering institutions which function like stock markets crossed with bookmakers: people trade stakes in the outcomes of future events. They mention that the Iowa Electronic Markets were forecasting a Bush victory in the 2004 election with 50-52% of the popular vote (he won with 51%) for a week beforehand, while many polls were still predicting a Kerry win even the night before. While the major polls have been accurate to 1.93% in predicting US election outcomes in the last 12 years, predictive markets have beaten that to 1.58%. James Surowiecki’s The Wisdom of Crowds explores this further.

The superior accuracy of predictive markets over traditional polls is quite simple to understand. Predictive market traders must put their money where their mouths are, whereas the polled must not, meaning that they frequently don’t reveal what what they really think. This is a great advantage of markets: the price mechanism allows us to see true subjective values.

Consider that with healthcare, which in Canada lacks a price mechanism, consumers will demand the moon on a stick: faster service, more equipment, more doctors, more nurses, more of everything. But when people are forced to buy healthcare with their own money, you will quickly find what people actually want most from their healthcare system and what they are prepared to sacrifice. The government has tabled some sort of plan to bring broadband internet to “all Canadian communities.” Imagine if this became truly universal and broadband was provided to all with tax money! People would demand faster and better broadband access and proclaim their right to it, even though many might not even want it at market prices. Right now, though, internet access is sold according to demand and subjective value. Some people who just want e-mail are happy with dial-up. Casual surfers are satisfied with a cheap DSL connection. Heavy internet users are free to purchase more expensive, high-bandwidth options. This, as we can see, lets people choose an internet package that they truly value, rather than asking them to assess the value of something that is essentially free. Of course, a person will demand any free product or service so long as its value is greater than nothing, and although you can find out if he values something, you can only accurately gauge how much he values it with a price mechanism in a free market.