Aggregation of multiple prior opinions (Q654529)

The authors consider a decision maker who is confronted with \(n\) experts having different prior sets of beliefs. There is a set \(\Lambda\) of weight vectors over the experts. The decision maker evaluates each act according to the maximin expected utility (MEU) model by Gilboa and Schmeidler. The experts adopt the decision maker's utility function (!). The authors prove the equivalence of two evaluation procedures: 1) The decision maker knows the sets of prior beliefs \(C_i\) of the experts and evaluates an act according to MEU with a set of probabilities derived from \(\Lambda\) and \(C_i\). 2) The experts only report their minimal (bottom-line) evaluation of an act and the decision maker uses \(\Lambda\) and these evaluations for the MEU-evaluation. Moreover the authors prove that these are equivalent to a condition called `Expert Uncertainty Aversion' (EUA). If each expert evaluates an act above the weighted average of the evaluations of others acts, so should the decision maker.