Foresight: The International Journal of Applied Forecasting, Issue 45, Spring 2017, Pages 38 – 42
The 2016 U.S. presidential election was a particularly bad case for prediction markets, as was the Brexit vote in the UK. In theory, these markets should be very effective in aggregating the information of individual forecasters into an overall market forecast. Because the individual participants must put “skin in the game,” they are expected to be more diligent about making use of relevant information than participants in surveys who are simply asked what they think will happen. In this article, Foresight’s Prediction Markets Editor Andreas Graefe reviews the recent performance of prediction markets to explain why the theoretical benefits of this approach to forecasting have not always stood up in practice. He raises the possibilities of market manipulation, participant misunderstanding, and bettors’ systematic bias.