Essays on decision biases in prediction markets

Many fields find predicting uncertain future events challenging. Prediction markets – futures markets around uncertain events – have become a major prediction tool. However, there is a lack of research on their calibration and decision biases by traders in prediction markets. In this thesis, study compensatory (Optimistic or Pessimistic bias), and non-compensatory (Favorite longshot or Reverse favorite longshot bias) in prediction markets and individual forecasts. We develop a statistical methodology to detect these biases among market prices and individual forecasts. We show both prediction market prices and individuals show both compensatory and non-compensatory calibration biases in their probability estimations. In a separate set of studies, we document decision making biases in political prediction markets. By analyzing trading justification in markets for U.S. and U.K. elections, we find provide strong evidence of emotional biases among traders. These emotional biases depend on whether the trader’s action supports or opposes the political status quo. Our text analyses show the tendency to use more positive (negative) emotional words when individuals justify their action of supporting (opposing) the political status quo.