The Elephant NOT in the Room
Prediction markets, as we learned in class, are places where people can trade securities that are tied to the outcome of some event, where if the event occurs, security-holders will earn a dollar, and otherwise, earn nothing. This means that the prices of securities are indicators of the market’s opinion on the probability of the event. The higher the price, the more likely the event – most of the time. On one such prediction market, Predictit.org, Donald Trump winning the election had about a 24 cent average, which means that buyers in the market thought he had about a 1 in 4 chance of winning. So what went wrong?
The key idea here is that the buyers in the market did not think he had enough support to win. However, this does not mean that these supporters did not exist. Many Trump supporters, believing his assertion that the media was rigged against him, did not participate in polls. This led to a vastly understated probability of him winning, an event that took us all by surprise. Prediction markets also failed to predict Brexit, Great Britain’s surprise vote to exit the EU, by the same idea. As this article claims, this could be evidence for the poor reliability of prediction markets as accurate predictors. With such a low payout/risk, the markets do not provide sufficient motivation for the buyers to be accurately informed. Also, the prediction market prices are self sustaining; this means that buyers take indication on their bid not only from their beliefs on the outcome, but on the current market price. This results in a less informed market. As such, it would be wise to take the results of a prediction market with a grain of salt, as they may be ignorant of the elephant (not) in the room.
https://www.bloomberg.com/view/articles/2016-11-15/prediction-markets-didn-t-call-trump-s-win-either