Uncertainty Attitudes, Subjective Beliefs and Decisions under Uncertainty

We conducted survey experiments to gather new data on uncertainty attitudes and subjective beliefs. Our survey targets a general population, focusing on health and financial decisions. We collect data on (i) individual attitudes toward risk and ambiguity, including individual-specific probability weighting and loss aversion and (ii) individual probabilistic beliefs about events related to the decisions we consider. The joint availability of individual preference parameters and subjective beliefs allows us to study heterogeneity and cross-correlations across these parameters in depth and to estimate structural models of decision-making under uncertainty that go beyond subjective expected utility. We find that both preferences and beliefs are highly heterogeneous. Compared to subjective expected utility, prospect theory combined with source-dependent probability weighting, fits the data strictly better in some contexts and equally well in others. Prospect theory performs particularly well when the belief distribution is concentrated towards extreme values.