Learning and Updating with Non-Additive Beliefs
Zusammenfassung der Projektergebnisse
Economic decision making usually takes place under conditions of uncertainty about the decision context. In general, many variables, which are known to influence the outcome of the action chosen, will be unknown at the point of decision making. Ambiguity describes a situation where such uncertainty cannot be represented by a (subjective) probability distribution. Over the last two decades, many differing approaches for modelling decision making under ambiguity have been proposed. All of them suffer from unresolved questions. Among these, questions about the impact of additional information on beliefs and subsequent choices are particularly important for economic applications. How will beliefs be revised in the light of new information? And how will these updated beliefs affect choices? This research project dealt with theoretical and empirical issues of updating rules. In particular, it explored the advantages and disadvantages of several updating rules in the context of economic models. We can show that Generalised Bayesian Updating of neo-additive beliefs has the attractive feature that the updated beliefs maintain the structure ofthe prior beliefs which is characterised by • a focal probability distribution, • a degree of confidence in regard to the focal probability, and • the decision maker's attitude towards this ambiguity, i.e., optimism and pessimism. In particular, ambiguity attitudes remain unaffected by the new information, while the focal probability is updated in the usual Bayesian way. We can demonstrate that it is possible to characterise both the Generalised Bayesian Updating rule and the specific form of neo-additive beliefs by assumptions on the decision maker's choice behaviour. This result is important because it opens up the possibility of testing the validity of this model of decision making under ambiguity in laboratory experiments.
Projektbezogene Publikationen (Auswahl)
- (2009), "A Dynamic Eilsberg Urn Experiment". Games and Economic Behavior
Peter Dorsch, Adam Dominiak & Jean-Philippe Lefort
- (2009), "Optimism and Pessimism in Games". Journal of Economic Theory
Jürgen Eichberger & David Kelsey
- (2009), "Attitudes towards Randomization: An Experimental Study". Economic Theory (Special issue in IHonorof Daniel Ellsberg edited by Mark Machina)
Adam Dominiak & Wendelin Schnedler
- (2010). "Agreeing to Disagree under Ambiguity"
Adam Dominiak & Jean-Philippe Lefort
- (2010). "Comparing Three Ways to Update Choquet Beliefs", Economics Letters 107, 91-94
Jürgen Eichberger, Simon Grant, David Kelsey
- (2009), "Unambiguous Events and Dynamic Choquet Preferences", Economic Theory
Adam Dominiak & Jean-Philippe Lefort
