Abstract There is a large theoretical literature in both economics and psychology on decision making under ambiguity (as distinct from risk) and many preference functionals proposed in this literature for describing behaviour in such contexts. However, the empirical literature is scarce and largely confined to testing between various proposed functionals. Using a new design, in which we create genuine ambiguity in the laboratory and can control the amount of ambiguity, we generate data which enables us to estimate several of the proposed preference functionals. In particular, we fit Subjective Expected Utility, Prospect Theory, Choquet Expected Utility, Maximin, Maximax, and Minimum Regret preference functionals, and examine how the fit changes when we vary the ambiguity. We find that the Choquet formulation performs best overall, though it is clear that different decision makers have different functionals. We also identify new decision rules which are not explicitly modelled in the literature.
Choquet OK?
G. LOTITO;MAFFIOLETTI, Anna
2007-01-01
Abstract
Abstract There is a large theoretical literature in both economics and psychology on decision making under ambiguity (as distinct from risk) and many preference functionals proposed in this literature for describing behaviour in such contexts. However, the empirical literature is scarce and largely confined to testing between various proposed functionals. Using a new design, in which we create genuine ambiguity in the laboratory and can control the amount of ambiguity, we generate data which enables us to estimate several of the proposed preference functionals. In particular, we fit Subjective Expected Utility, Prospect Theory, Choquet Expected Utility, Maximin, Maximax, and Minimum Regret preference functionals, and examine how the fit changes when we vary the ambiguity. We find that the Choquet formulation performs best overall, though it is clear that different decision makers have different functionals. We also identify new decision rules which are not explicitly modelled in the literature.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.