Analyzing the Analyst. Heuristics and Biases, Group Descision-Making and Rational Herding in Forecasting Experiments
The behavior of analysts doing economic forecasting has been discussed in a number of empirical studies with the overall result of a very limited quality of their predictions. Further, a strong correlation of forecasts and current values, as well as herding behavior of analysts is shown. These findings have been confirmed for various economic variables. However, the reasons for this pattern of analysts? forecasts remain unclear. The seven chapters of this dissertation thus investigate the finding of poor quality of forecasts, particularly regarding their correlation to current values, using the methods of experimental economics. Potential influences on analysts? forecasts are considered for individual decision-making and within social contexts. Regarding individual influences on forecasts, the anchoring bias is investigated within an experimental forecasting setting. Two ensuing papers implement the role of endogenously derived anchor values as well as the effect of anchoring within group cooperation. Further papers deal with the effect of the mode of communication on group rationality, as well as with the overconfidence bias in the domain of forecasting. With regard to rational herding, an experimental study compares informational efficiency and welfare effects of endogenous and exogenous timing of decisions. A further experimental study integrates various potential influences on forecasting behavior in a single experimental design and thus shows the coordinative effect of reputational herding within a forecasting setting.
Period of Dissertation
2011 - 2014