We propose second-order belief model of decision-making under ambiguity, that extends the Salience Model of choice under risk (Bordalo et al., 2012). Our model is able to predict stylized facts of the literature, such as likelihood insensitivity, the fact that people tend to overestimate(underestimate) the expected probability of low(high)-likelihood events to happen. We also predict the fourfold pattern of ambiguity attitudes, where decision-makers are usually ambiguity averse for bets on high probability gains and low probability losses, and ambiguity seeking otherwise. A key feature of our model is that ambiguity attitude is a result of the combination of: (i) the outcome domain (gain or loss) of a bet; (ii) the expected probability of the events that are relevant to the outcome of the bet; (iii) how many states of the world are deemed possible by the Decision-Maker. So, there is no need for additional assumptions on our model to predict that ambiguity attitudes change according to those three factors, in agreement with the experimental and empirical literature. Furthermore, we include the context - represented by characteristics of the choice set of the decision-making problem - as a variable that may affect preferences, in the same spirit of the original Salience Theory.
Comissão Organizadora
Anderson Odias da Silva
Claudia Yoshinaga
Ricardo D. Brito
Felipe Saraiva Iachan
Vinicius Augusto Brunassi Silva
Complete tool for submission, evaluation and publication of proceedings for scientific events.
