Regional Economic Bulletin J U N E 2 0 0 5 Prediction Markets: the Virtual

Predicting the future is a tricky business. In hindsight, books that list expert predictions prove to be utterly worthless.2 Yet all finance choices, from deciding to introduce a new product to which stocks to purchase, depend on future uncertainty. A major role of financial markets is to assess the likelihood of different future events occurring, as markets try to predict earnings, interest rates and many other economic and political variables to arrive at financial prices. This article examines a recent innovation in finance: internet prediction markets. These markets, which have evolved from gambling markets, specialise in predicting actual events rather than prices of financial securities, as financial futures and options markets do. Many types of events, from election results to the identity of the new pope, have been traded as contracts on prediction markets. The internet, with its low costs and ease of trading, has been the innovation fueling the growth of prediction markets. Two prediction markets are analysed that demonstrate their abilities and usefulness: the New Zealand Virtual Super 12 and Tradesports' contracts on the US presidential election. Why do prediction markets perform so well? People that follow the stock market are well aware of the predictive power of stock markets. For example, when the space shuttle Discovery exploded in 1986, one of the major contractors, Morton Thiokol, experienced a major 12 % decline in stock price that day. The other major contractors had only minor losses of less than 3%. As it turned out, the stock market had correctly assessed who was to blame months before NASA was able to determine the cause